Jonah Kowall

A member of the Gartner Blog Network

Jonah Kowall
Research Vice President
3.5 years with Gartner
20 years IT industry

Jonah Kowall is a research Vice President in Gartner's IT Operations Research group. He focuses on application performance monitoring (APM), Unified Monitoring, Network Performance Monitoring and Diagnostics (NPMD), Infrastructure Performance Monitoring (IPM), IT Operations Analytics (ITOA), and general application and infrastructure availability and performance monitoring technologies. Read Full Bio

Coverage Areas:

Got Nagios? Get rid of it.

by Jonah Kowall  |  February 22, 2013  |  169 Comments

Nagios is a great product, it’s free, you can’t beat that. The problem is that the level of usability and sophistication of the product is pretty much zero. Don’t expect any bells and whistles, or really any usability for that matter. The technology is rudimentary at best, but it can get the job done with the right skills on staff.

Many vendors have introduced products which make Nagios more usable, these improve the product itself, the supportability, and the fact that you can get support when things break. The problem is that the underpinning and ugliness still exist once you get through the layers intended to cover up the mess that Nagios is. There are still scripted “checks” which run to determine service health, the checks are normally challenging to manage, especially when some execute through the agent, while others do not. Other features added include better management, dash boarding, and other basic capabilities that you would expect out of the box with any monitoring product.

The problem with all of these approaches is that they don’t auto-configure themselves, they don’t detect application instances properly or consistently, and configuration of checks is painful. Most clients using Nagios will hear me tell them to ditch it, and go for a simple and inexpensive monitoring tool. I hear from many Gartner clients who decide to implement open source tools based on a talented engineer on the team, but when he leaves the company no one can figure out how to safely upgrade nagios or it’s associated components (This article goes through some of what is needed to manage Nagios : http://www.debianhelp.co.uk/nagiosinstall.htm)

The time and effort needed to manage this software is much better spent buying a simple monitoring tool to get the basics covered for infrastructure health. Once you lick the easy stuff, infrastructure health monitoring, you can start focusing on the harder problems. Application performance monitoring (APM) tools are where most interest is since they facilitate end user experience monitoring, in depth troubleshooting capabilities, and provide much greater business value to the non-technical users.

[EDIT : 11/12/13]

Other Nagios related blog posts:

  1. Nagios : Let the religious wars continue

  2. How to properly leverage open source server monitoring (Nagios)

  3. Monitoring software sucks so I use Nagios, what’s a better approach?

Probably a better way to do similar types of monitoring, from a wider perspective:

Unified Monitoring – Note / Presentation and Client Interest

169 Comments »

Category: IT Operations Monitoring     Tags:

169 responses so far ↓

  • 1 Paul De Audney   February 22, 2013 at 8:43 pm

    Combine Nagios with a modern configuration management tool like Chef or Puppet and you can have your monitoring sensors automatically applied to your running Nagios configuration when you use either tool to build out your infrastructure when hosts are deployed.

    Additionally Nagios is widely used and it is easy to find people with significant experience using and optimizing it.

    While it is not the greatest monitoring tool out there, there are many add-ons that go a long way to make it significantly better than commercial offerings. Not to mention the integrations that can be achieved quite easily (shutting down a host for a scheduled reboot? Have the host notify the nagios server of the downtime while it is shutting its services down.)

    You can configure nagios with service dependencies so when a fault occurs inside your infrastructure, you only get alerts to the actual fault. This does take some thought, blindly following a guide designed to get a newbie setup with Nagios isn’t going to result in a good example.

    Have you ever actually configured nagios with any of these integrations? I am yet to see a commercial tool with the flexibility that can be achieved with nagios and a little bit of scripting and experience in current system administration practices.

  • 2 Jonah Kowall   February 22, 2013 at 9:29 pm

    While I agree that it can be effective in controlled smaller environments this is not the situation in most enterprises. The skills needed to set it up, manage it, and build that level of standardization and control is not normally feasible. Each admin sets it up differently in terms of the scripts used, not unlike puppet patterns or chef recepies. This is the root of the problem, the sysadmin over customizes the monitoring to do things better suited for other tools, hence over extending a product which has scalability issues.

    You can configure dependencies which are single tree, but that’s unrealistic as well since we have multiple levels of dependencies which modern monitoring tools will detect automatically especially from the network topology, or virtualization dependencies.

    I have configured Nagios before, and even in production environments, but it was easily replaced with much more simplistic tools for small amounts of money.

  • 3 James   February 25, 2013 at 3:24 pm

    this depends on what you consider “small amounts of money”. Being in the industry I know there is virtually no such thing. At the Nagios Conference last year it was also revealed one of the largest insurance companies in the USA uses Nagios. They have it monitoring 1.25 million devices and it doing it well. There was another person monitoring million as well. This article is highly inaccurate and has nothing to offer fact wise.

  • 4 James   February 25, 2013 at 3:25 pm

    You can also get support for free via Nagios Ent. or purchase a support contract so it is supported if you want it to be. FYI. Fact checking is a good thing before writing an article.

  • 5 Jonah Kowall   February 25, 2013 at 3:38 pm

    You can get good monitoring products for anywhere from $2-$20 per device (at small scale) that work well enough, they are not sophisticated but functional and automated.

    Nagios with agent architecture (SNMP can scale better) does not scale, the check architecture just cannot handle volumes. I’m sure the large implementations have done the following:

    1. not using agents with checks
    2. changed the code to optimize it
    3. are using a lot of hardware to scale the solution

    If you want to dispute my facts, I speak with well over 100 clients per year using Nagios, this is what I hear from clients. These are clearly issues with the product, as well as maintaining the product after the engineer with expertise leaves the company. I don’t just hear these things a handful of times, they come up regularly.

    I wrote a detailed note a year ago outlining support options for Nagios, not to mention there have been modifications done to the product to help correct some of these issues by those companies. Regardless the product is still pretty poor, everyone agrees with that statement.

    Thanks for the comments, and nice job hiding your identity :)

  • 6 James   February 25, 2013 at 4:31 pm

    With Millions of users world wide you speaking in absolutes about what every one agrees with is a pretty poor statement for an analyst. The agents do scale better than SNMP. Nagios XI has the ability to monitor via SNMP and WMI. The agents scale far better. Nagios Core, what your talking about also has the ability to monitor via SNMP. Again, you should fact check before making statements in absolutes.
    Maybe your talking to the wrong people. If you talked to MTV, Mcafee, Toshiba, Johns Hopkins, Universal, Yahoo, Siemens, Sony and much more they might tell you different.

  • 7 Ethan   February 25, 2013 at 8:32 pm

    Are you referring to Nagios XI, Nagios Fusion, Nagios Core, or something else altogether? As the research director, I would have thought you’d investigate what you write about, rather than making broad-reaching statements that come off as uninformed to the knowledgeable IT person.

    Nagios Core by itself may not be terribly user-friendly to non-technical folks, but there are thousands of addons developed by the community that make it work well. Nagios XI has an extremely intuitive interface and makes configuration (and auto-discovery) easy.

  • 8 James E   February 25, 2013 at 9:13 pm

    You really need to do some more research.

    Nagios XI is EXTREMELY easy to use and configure. Actually the latest version of Nagios XI pretty much negates all arguments you put forth.

  • 9 Phil   February 25, 2013 at 9:14 pm

    What a load of piffle!

    We use Nagios to monitor thousands of services on around 500 hosts without any problems.

    The ability to write and deploy one’s own plugins put it streets ahead of the dumbed-down point & click windows products, which never quite seem to do what I want.

  • 10 PapaLinux   February 25, 2013 at 9:22 pm

    I will resume this comment, Nagios is:

    Flexible, avalible for being installed in any linux distribution from repositories and from source code, you can develop your own plugins, you can obtain information by snmp, wmi, agents. There are a lot of plugins and addons available at nagios website ready for being implemented.

    And if we talk about Nagios XI, any guy with a simple intuition can configure any device for being monitored.

  • 11 James E   February 25, 2013 at 9:24 pm

    I also enjoy how you use the term “geeky Linux guy”. You are basically writing this article to a bunch of windows admins who have no concept of how to configure a host check beyond clicking next, next, finish. (not all are that way)

    Please stop basing your articles on information from 10 years ago just in case someone actually takes you seriously.

  • 12 Rodrigo   February 25, 2013 at 9:38 pm

    I dont understand, I have two customers that use 10000 checks with Nagios XI. First version that they used and not ask help to use Nagios XI. Is very easy. I don’t Agree.

  • 13 Luka   February 25, 2013 at 11:10 pm

    My company is a Nagios reseller and integrator and we have around 10 Nagios implementations so-far.

    We have done SWOT analysis of various monitoring software in terms of performance, price, features and other categories, and Nagios scored very high (second).

    Although I definitely agree it’s a “hackish” software, at least in it’s Core version, it’s extremely useful, flexible and does not take a huge amount of administration and knowledge. Typical Unix admin can get the basics in a week for sure…
    Some of your points are invalid, but it is obvious that you’re referring to Nagios Core, not Nagios XI (btw, I like Nagios Core more)…

    We have couple of 200+ hosts and 2000+ services implementations with *minor* tweaks and they run perfectly on a 4-core VM!

    It’s true, however, that Nagios Inc. should invest more time not on usability only, but in being an “enterprise-accepted” company. Not having *official* Windows agent (NSClient is not official, it’s third party), necessity of compiling packages for virtually every Unix out there (please, create *official* and approved packages and versions) and some minor issues with Nagios XI (bugs, immaturity) give impression of immature and “hack-your-own-tool” software.

  • 14 Kevin J   February 25, 2013 at 11:59 pm

    This is why I don’t trust Gartner! They do publish good research if you pay for a membership – again if you pay and can afford to. I’ve installed Nagios Core and used NConf to configure it in dozens of locations cheaper and more scalable than most other tool that claim to be good enterprise tools for a decent price with a per agent cost by the way. One instance I setup watches over 100K services with 2 minute checks on a small VM. A good “geeky Linux guy” can make a wimpy VW Bug into a jet powered racer with ease and there are plenty of us out there!!! The custom scripting to extend monitors make the system not only robust to be able to handle custom scripts but also let’s you quickly extend functionality without waiting for vendors to do so.

  • 15 Mike   February 26, 2013 at 12:13 am

    Hi Luka, a windows version of XI will happen eventually I would think. We are still a young company and have plenty of growing to do. NsClient is a community project though it is a very good nagios project that Nagios Ent. provides funds for. NsClient I believe was also originally created by Ethan (Nagios creator) and is now maintained by a well known Nagios community member. As a partner you will be notified if any of this changes. And it could soon :)

  • 16 Mike   February 26, 2013 at 12:20 am

    I too am saddened by the quality of this post. I would think that the director of IT of all people would know this was poor quality. Maybe that’s the secret to success. If you want to be the most widely used monitoring solution it has to be poor quality. If you want to be a very well known analyst firm you need to post articles of poor quality. According to Gartner any way.

  • 17 Jonah Kowall   February 26, 2013 at 1:19 am

    James : I have requested briefings on Nagios XI several times, and I have not been briefed properly. I have also spoken with happy and unhappy customers of yours.

    Mike : The trick to being widely used is being free and extensible, the issue is that most users of the products agree the product is quite poor and could use major improvements. There aren’t any stellar monitoring tools out there from a general health monitoring perspective, so why bother paying for it? Someone will fix this and get the proper product on the market, but today for system monitoring it doesn’t exist.

    If you trust Gartner or not, we have the most conversations with builders and implementers of IT products, hence we have a wider viewpoint than other 3rd parties have. You can agree or disagree with my opinions, I’m not going to debate you on a technology war especially not in public, since frankly it’s not constructive. The solution for one situation is not the same for other situations, hence providing customized advice is the value.

    These statements are around Nagios core, and not the commercial offerings which extend the core processing functionality.

  • 18 Johannes Dagemark   February 26, 2013 at 12:16 pm

    Constructive critisism is always a good thing, taking about things one do not understand, well perhaps not so much.

    If I were a windows admin on a small company I would not use Nagios either, there are plenty of good tools that would take you to an acceptable level of monitoring in a much more userfriendly and easy way.

    But for medium to large companies with linux competence inhouse I have yet to find a monitoring tool that flexible, feature complete and most important of all as high performing that Nagios is.

    Do you know of another monitoring tool that can monitor 1.25 milion devices? I dont..

  • 19 Ashley   February 26, 2013 at 3:24 pm

    Wonder how many of these comments are from Nagios employees/resellers?

  • 20 James   February 26, 2013 at 3:45 pm

    I appreciate your comments Jonah. You haven’t asked for a briefing though, FYI. I’ve quoted you below:

    The management team at Nagios doesn’t really speak with analysts or really do any effective marketing, It’s of no cost for a vendor to brief analysts, and all of them should do so. Normally I request briefings when clients ask about products, but I haven’t had a client ask about XI yet!

    We had RedMonk at our conference as well. They wrote positively on our behalf several times now.

  • 21 Luis   February 27, 2013 at 12:06 am

    Hi,

    I’ve being using Nagios Core and XI, they are both great products, even I have compared them with other solutions, and it’s amazing how flexible it’s Nagios. Being objective, before deploy a ANY TOOL read the instructions, it’s not good to say that the product is bad if you don’t know how to implement it ;-) .

    If you a Windows Admin, I really know how you would feel trying to make work a solution like Nagios on Linux ;-) , remember if a tool offers me a lot of stuff like Nagios, I stay with Nagios.

  • 22 Jonah Kowall   February 27, 2013 at 1:49 am

    James, I have 3 emails proving I did in fact ask for briefings! Please set one up.

    http://www.gartner.com/technology/about/vendor_briefings.jsp

  • 23 Raul   February 28, 2013 at 9:57 am

    I think that the problem with Nagios is even worse.

    Even if it were as scalable, flexible and usable as some commenters claim, the problem with Nagios is that the aim of the product is to monitor devices instead of business services.

    This clearly limits the value users obtain from it. And it makes harder to evolve (I know this for experience) once the IT department matures and decides to go to the next level, id est, BSM.

    Nagios (and other similar solutions, for free or not) makes IT the center (inside-out approach) when the business should be the core (outside-in approach, see Ian Clayton blog which is great) of a monitoring project.

    Of course, IT geeks will be delighted with Nagios approach and would never understand this until it is late. That’s why CIO involvement is required.

    In any case, we at Tango/04 have done “upgrades” of Nagios and in other cases we just put us on top of it as part of a BSM project.

  • 24 James   February 28, 2013 at 4:45 pm

    Raul – You should look at Nagios’ BPI component. It’s built for business process monitoring. BPI stands for Business Process Intelligence. Nagios Core is supposed to be an engine for monitoring. The adons and components are what make it what it is. There are a ton of different front ends out there. I’m confused… It’s so horrible that you use it as your internal engine?

  • 25 Raul   February 28, 2013 at 8:56 pm

    I don’t use, some customers of us use it and keep it and some others prefer to just replace it completely. We may sit on top of Nagios or any other infrastructure monitoring solution, or just use our stuff.

    I agree that there could be a lot of front ends out there, and I suppose it is a good thing. In that regard, even Tango/04 could be considered kind of a front end.

    Still, in my experience Nagios is being use widely for infrastructure monitoring mostly. Inside-out approach.

    And I never said Nagios is horrible, James.

  • 26 Rodney   March 6, 2013 at 1:41 am

    This kind of article is from people that have never worked with a IT monitoring tool. People that has never support their clients with fast and cheap solutions. People who think that the more expencive the better solution, because in this way they seem more important for the company, as they are responsible for expensive budgets. People that are not worried to do a good job, but that want just appear to the others.

    I worked with monitoring tools for almost 20 years, and I consider Nagios one of the best, mainly because its flexibility. I´ve already worked with a lot of commercial tools and all of them, with no exception, are complex, inflexible, expensives, with a lot of bugs and bring almost no return to the company. For all of them we must hire a team of experts, what make the solution more expensive.

    I´ve never believe in Gartener articles or recomendations and after this articles I believe less than ever.

    P.S.: I´m not a Nagios employee or reseller.

  • 27 Jonah Kowall   March 6, 2013 at 1:15 pm

    Rodney, I’ve actually implemented well over 100 tools in production environments and have worked in IT for 17 years as an end user (such as yourself). While I agree Nagios is cheap, it’s not \fast\ compared to many lighter-weight tools. I disagree with you on the point of expensive = good, and normally I recommend easy tools to clients.

    Aside from my end user experiences I speak with well over 600 end user client organizations who are Gartner clients. I hear these comments overwhelmingly from them. They are normally using Nagios due to cost, but it doesn’t provide what they need. Many of them realize that the cost of maintaining and managing the product is higher than easier tools on the market.

    On another note there have been some nasty NRPE exploits circulating on some message boards, I would patch!

    http://packetstormsecurity.com/files/120507/Nagios-NRPE-2.13-Code-Execution.html

  • 28 Ross Fomerand   March 13, 2013 at 8:01 pm

    “These statements are around Nagios core, and not the commercial offerings which extend the core processing functionality”

    Glad you were able to clear that up for your readers AFTER the fact in the comments section. How about you edit the article now to include exactly which nagios product you are referring to before you lose even more credibility.

    Since your focus seems to be on open source monitoring technologies; I expect to see an article soon about how crippled the Zenoss open source version is out of the box. Looking forward to it.

  • 29 C. Mansfeldt   March 18, 2013 at 8:54 am

    “If you don’t take my advice I will probably be talking to them 3 years later when the “geeky Linux guy” leaves the company and no one can figure out how to upgrade nagios”

    I am not sure where your hate or anger is coming from, but please don’t denigrate an entire group of professionals because of it.

    If you had a shred of validity in your argument, you lost all credibility when you insulted the people who work hard to bring a solution to our employers that costs them nothing. Good job in alienating whatever audience you may hope to gain.

    -CM

  • 30 Jonah Kowall   March 19, 2013 at 2:40 pm

    Thanks for the comment Ross, I actually see more successful customers using Zenoss core (open source) versus Nagios (open source). There are commercial offerings on both camps which improve the situation.

  • 31 Jonah Kowall   March 19, 2013 at 2:55 pm

    No hate or anger here C. Mansfeldt, I speak to clients using Nagios who are happy as well, but I do get many calls from clients who’ve ended up with an unsupportable, and insecure monitoring system. If you don’t keep the systems and packages up to date it moves towards disrepair.

  • 32 Eric Anderson   March 28, 2013 at 3:43 pm

    We’ve seen a lot of migrations from Nagios to CopperEgg as people move into more dynamic environments (like public/private clouds). I think in general people are tired of the old clumsy monitoring tools and ready for a next-generation tool that looks and feels the way the rest of their applications and tools work.

  • 33 CopperEgg vs Nagios: Installation guide, maintenance and price   April 3, 2013 at 4:06 pm

    [...] seen here in a Gartner blog post: “Most clients using Nagios [Core] will hear me tell them to ditch it, and go for a simple and [...]

  • 34 Andreas Ericsson   April 4, 2013 at 7:00 pm

    Hi there. I’m one of the current maintainers. I work for a company reselling Nagios-based solutions.

    As for usability of basic Nagios core, I must say you’re right. It does have quite a steep learning curve. For vendors, that’s awesome though, because we can easily bridge the gap so anyone can (and do) manage it.

    As for performance, I can only assume you haven’t tested the latest and greatest, or that you failed really hard in setting it up. My test-configuration a quarter of a million points of measurement (“services”, in nagios terms). I run that on my laptop for several hours every day, with no latency, having Nagios eat somewhere between 8-15% of my cpu and about 250mb of ram. Watching a youtube clip uses about the same resources.

    You’re also talking about “service level management” and how Nagios lacks it, which tells me quite clearly that you’ve either never worked in a large enough network, or that it was years and years ago since you did things like that. I’m guessing you’ve forgot the roles in an organisation that benefit from monitoring. I’ll run them down for you here, as I’ve done in dozens of presentations over the years.

    First off, there’s the admins. They need to know within a reasonable timeframe (ie, as soon as possible) when something breaks down. If they don’t, they can’t do their job, and the company grinds to a halt before someone notices that the inhouse thingumabob has gone AWOL. Their only reason for knowing which (user-level) services are affected is so they can prioritize what to do first when too large a pile of crap hits the fan.

    Second, there’s the internal IT support staff. They really need to know *why* people can’t send email, and when that’s intended to be fixed. They really don’t care if it’s because a router is sending all packets to the printer or because the DNS is down; They just want something to say on the phone.

    Third, there’s the CIO and CTO. They really want to know the availability of the (user-level) services so they can present shiny graphs to their bosses so they can prove that their hard work provides improvements so they can get their annual bonuses.

    Fourth, and last, there’s the CEO and CFO, who just want to know exacly how much money everything’s costing and which contracts they can eliminate without it costing them more than they save. They like hard numbers on that sort of thing. And graphs. Everyone who is too incompetent at a real craft to wear what they like to work enjoys looking at shiny graphics, it seems.

    The point is that the only way you can achieve all of the above is to start from the ground up and then aggregate information for whoever’s looking at it. But in order to start from the ground up, the tool doing the watching has to know what to watch, in as much detail as you want to be able to read it.

    Given how obvious it is that computers should support humans rather than the other way around, and given how important all of the above humans are to any IT-based organisation, I have a hard time understanding where your rant comes from.

    Oh, and for comparison; Paying 2-20 dollars per device you want to monitor (what’s that for btw? per month? per year?) is quite a lot of money compared to what it would cost to hire a consultant to set up Nagios for you.

    As for huge companies monitoring their huge networks; No, they don’t modify Nagios’ code. NASA doesn’t sit down and hack around in Nagios’ core to watch their 10k-node supercomputer clusters. They do stay in touch with the Nagios developers to figure out how to make the most of it though, and they investigate addons designed specifically to increase robustness and performance. In comparison to (insert random database here), people spend significantly *less* time getting Nagios to perform than they do on (insert same database here).

    So, while you do have some valid points regarding usability, most of your essay seems to be taken out of thin air, or done as a weekend project by a fifth-grader. I would really have expected a grown man who claims to have been an IT professional for 17 years to do their homework better.

  • 35 Jonah Kowall   April 4, 2013 at 7:10 pm

    Andreas, I believe that when you add in technologies such as the vendor you work for that the solution becomes infinitely more manageable and handles a lot of the downsides within the open source technology itself. The agents used by Nagios are still lacking configuration management, proper policy management, and patch management. This means you must either use additional tools to manage those elements or buy a commercial product which helps handle the shortcomings within the architecture.

    I have implemented your product, open source products, and other similar Nagios based solutions.

    When speaking about SLA management, it’s not just about managing SLAs (which Nagios does in fact do) its about managing, correlating, and understanding topologies of systems, networks, and other elements (storage, virtualization, etc). While your product handles those, the open source product lacks topology understand and event management across those elements.

    If you are claiming that monitoring tools that do not provide APM technologies (such as Nagios) are relevant to the C level executives that’s clearly not the case, although the operations team will build these dashboards the metrics are meaningless without looking at actual end user experience.

    Most businesses do modify the check code (or even use various overlapping check code) in place, and create management issues, this is something I hear from clients regularly. While they aren’t modifying core Nagios code, they are modifying other parts of the codebase.

    I’m not responding to your personal attacks here, that’s simply false and non-constructive. Tell Jan I send my regards :)

  • 36 Micheal D jackson   April 26, 2013 at 5:17 am

    Agreed with the responses….couple it with any solid configuration management framework with automation scripts Nagios checks automatically deploy themselves!! I don’t think I’ve had to manually deploy checks in a year and I’ve deployed over 400 servers :-).

    Really it’s how you design your foundation from the ground….not just the tools themselves.

  • 37 Florian Heigl   May 6, 2013 at 11:39 pm

    I read your article and I pretty much wonder whether you have really followed the development in Nagios’ ecosystem after 2009.

    Much of the things you state don’t hold any more these days, especially w/re of autodetection / config management.

    I’ll just say Check_MK.
    I’ll not say that people drive their nagios config from NATIVE access to IPAM or CMDB with very light scripting. I’ll not say rule based configuration. I’ll not say anomaly alerting. I’ll not say trend monitoring.
    I’ll say that that’s just one of multiple projects that are moving the Nagios concept where it wasn’t expected to go.

    I’ve seen Nagios instances that ONLY monitor business KPIs instead of technical infrastructure. It’s a matter of creativity and flexibility, and Naigos has enough of the latter.

    Yes, plain Nagios is stuck where it was in 2003, but who cares anymore? It’s 2013 and we’re far beyond that point.

    Sorry, “Gartner”, but you slept at the wheel.

  • 38 Michael   May 16, 2013 at 4:52 am

    We monitor clinical health system enterprise solutions for hundreds of hospitals across the US using simply Nagios core. It’s far superior to any clinical vendor software that I have seen, highly flexible and allows for business intelligence alert capabilities. You just have to understand the basic framework and build it properly. From that point on, expanding the solution or adding additional services is extremely easy to run, manage, and deploy. If you’re looking for a solid monitoring platform and you happen to read this article, don’t be fooled!

  • 39 Todd   May 18, 2013 at 4:09 pm

    We have been using Nagios core for 5 years to monitor our data center. Before we used several other solutions but got tired of the high price and lack of flexibility. You could monitor services or hardware setting, but they didn’t allow you to monitor both a Postfix mail queue and a Windows mail queue, a MySQL server and SQL. They didn’t allow you set custom alerts based on the server and services. We can create custom logons for our clients so they only see their servers. The list goes on….

    Nagios is without a doubt one of the most flexible monitoring solutions available.

    It is Stable, require few resources, the web interface is easy to understand though not glamorous. And it has NEVER broken in 5 yrs.

    Ask yourself one question. Why would I pay for a product and expect less?

    In the future do your research before you post articles.

    The simple fact that most every response to this article denies your “opinion” rather than thanks you for the info should is very telling.

    NOTE: I am not a programer or even a Linux guy, 90% Windows But I setup and run our SLES based Nagios and have maintained it with little effort including the updates for 5 years.

  • 40 Jonah Kowall   May 20, 2013 at 9:12 pm

    @florian – Indeed and I speak to Nagios users (happy and not daily). The fact that you are speaking about CMDBs shows you haven’t been following the industry trends away from such technologies, but they are in use in many enterprises with lackluster successes.

    @Michael – Agreed its flexible, and customizable, but the cost of doing so, upgrading it, maintaining it, and securing it are elusive to most shops without expertise on staff. Over time that staff often changes roles leaving the abandoned monitoring in place. This can result in compromised systems, and missed alerts (among various other issues) I hear from clients on a regular basis. It looks and works great on day 1, but after day 1000 the product is not always so.

    @todd there are many products which can do so without agents easily, Nagios is one choice, but it comes with risks. If you read carefully, I’m not saying get rid of all Nagios based technology, I’m saying the open source product has major support headaches. I alos notice as many others who have posted here that you enjoy slamming my “research” but I speak with more Nagios users that you can imagine, and having implemented open source and closed source monitoring technologies for 15 years before joining Gartner I have plenty of battleground experience (and scripting experience across multiple tools and languages). My opinion is based on thousands of discussions with clients, and dealing with hundreds of vendor solutions, hence my opinion is likely well informed. This is the reason clients come to Gartner for advise.

  • 41 Sam Bould   June 7, 2013 at 8:05 pm

    Hi Jonah,

    Does not the fact everybody disagrees with you about nagios make you reconsider your position, just a little bit? You dont like the tool but it doesnt mean you are right and the rest of profesional (100% or your comments are against you, including me as Nagios technical user) are wrong.

    I mentioned before in other blog entry that you cannot provide this non-supported by facts killing opinions. Does the tool have everything wrong?

  • 42 Jonah Kowall   June 7, 2013 at 9:14 pm

    Sam, actually most clients agree with me, but the folks posting on this blog are using it in specific environments which are not the typical enterprise client.

    I’m not saying Nagios is bad, I’m saying the open source tool ends up costing significantly more than the license price for most environments. You can read countless research articles talking about the cost of open source, sometimes it’s a good bet, and other times it is not. Nagios is one example where the open source isn’t a good bet.

    There are plenty of low cost Nagios based products which correct the shortcomings of the open source product. In fact companies like OP5 write most of the nagios code for the open source product! You wouldn’t have Nagios 4.0 (major rewrite) without these companies, so go and support them!

  • 43 Wayne   June 17, 2013 at 12:17 am

    As with all opinion based articles, you’ll always get arguments. I read very few facts in the original document.

    The fact that there are lots of monitoring products out there verifies (for me) that no one product meets everyone’s requirements. Where I work we use many monitoring tools – some commercial, others OpenSource or developed in house. Why, because no one piece of software does everything. Nagios is part of the picture and I can definitely say that none of the products come without additional staffing costs – they all require ongoing configuration changes for our environment.

  • 44 Cameron Fuller - Why should I use Operations Manager instead of open source monitoring solutions like Nagios? (#SCOM, #SYSCTR)   June 18, 2013 at 1:41 pm

    [...] http://blogs.gartner.com/jonah-kowall/2013/02/22/got-nagios-get-rid-of-it/ [...]

  • 45 Joshua   June 30, 2013 at 1:42 am

    Jonah, I work in a typical enterprise client. We would not agree with you. We are moving away from those so-called “enterprise” monitoring tools and using Nagios instead simply because of the flexibility and the community behind it.

    There’s no silver bullet or magical tool that does what everyone needs. We’ve tried them all and there were always issues. Nagios has it’s issues just like the others, but we decided it was better to stick to the one that was most flexible and one that you can find resources for.

    It’s much easier to find someone with Nagios experience than it is for other monitoring systems. Search for a Unix admin’s resume. Odds are, you’ll see Nagios on there. Now search for one of the other tools.

  • 46 Rob   July 8, 2013 at 9:39 pm

    I would stop blogging about Technology. Seriously. If you think Nagios is “rudimentary”, all you are showing is how little you know about the subject. You must be a Windows admin.

  • 47 Jonah Kowall   July 8, 2013 at 10:37 pm

    Nice rob. Its very rudimentary compared to other tools, talk to the main comiters they will agree!

  • 48 Luke   July 9, 2013 at 12:21 am

    This article demonstrates the authors complete lack of understanding of Nagios as a solution. Everything mentioned is complete nonsense, sounds like someone needs to do a Certificate 1 in IT…..

    All I got from this article was “I’m to stupid, was to hard, nagios is crap”

  • 49 Josh   July 9, 2013 at 2:34 pm

    Jonah,

    I have used things like Solarwinds IP monitor, MOM, and Zenoss, and to me I was very disappointed by those compared to Nagios. Yes I am a linux admin, but I think if anyone takes the time to learn how to use it they will be surprised. Nagios is very robust.

    It uses industry standard development languages. This means you don’t have to learn some specialized language that the monitoring system imposes on you.

    It is highly customizable so you can use what ever you want to monitor including your own bash, perl, python, vbscripts.

    Very lean. It uses very little resources in my opinion compared to others.

    I have used it in very small shops when I was consulting, I have used it in a healthcare environment, and I am using it in a 200 Server, 1000 end user environment.

    So in short I think I really disagree with your opinion.

  • 50 Lance   July 11, 2013 at 3:02 am

    The article says that you’re better off with a “simple monitoring tool” but fails to mention any. Care to share? I’ve worked with a lot of monitoring tools, most are not simple… and if they are, they’re useless.

  • 51 Davor   July 11, 2013 at 1:41 pm

    Yes, Nagios is a picture of the IT knowledge level at the organization, how it is implemented, used and interpreted

    i am since 25 years IT industry….

    P.S. I remember Gartner article about “Drop the Linux” to…

  • 52 Dan   July 16, 2013 at 9:45 am

    This article is hilarious.

  • 53 Jill   July 22, 2013 at 5:30 pm

    I am pleased to have read so many positive comments around Nagios.
    As a Nagios user (not a reseller, not getting paid in any way for saying nice things about the product), I think it is a fantastic solution.
    I have used several Gartner-recommended products (worked for companies that had CIO’s who placed too much value in Gartner) and I’ve found Nagios to be 1. easier to implement 2. easier to manage 3. easier to rollout AND 4. accurate.
    Nagios is a great FREE tool and I am sad to have read an article that dismissed it out of hand. There are hundreds of reasons to use Nagios and the author could have mentioned any one of them to paint a fairer and more realistic picture.
    The fact on Nagios – you can use it in your big or small environment to quickly provide return on your investment of time installing/configuring the tool.
    Don’t let this article scare you away from trying it out and making up your own mind – it IS worth your time to use a tool like Nagios.

  • 54 meh   July 25, 2013 at 12:28 pm

    I couldn’t agree more Nagios is such Garbage. Who has the time to waste editing files manually to configure hosts? Aren’t we past that as an industry? I guess Nagios resellers are flooding your comments because they sell support for this garbage product. I couldn’t agree more, stop wasting countless hours reading forums and editing config files to make it work and just get a good paid solution.

    Using this is like trying to piece a puzzle together that just doesn’t fit

  • 55 Manish Kumar   July 26, 2013 at 7:27 am

    I am amused to read this article by gartner, while I don’t know how gartner evaluates a product or simply they speak positive if big vendors are their partners or clients, I sense that nagios would have refused to be their client though for pitching their sales which they don’t need.. :) as it’s well accepted and has large installation base for its opens source version nagios core and the commercial version nagiosXI also doing good businesses as per their reports. However still companies can bet high on it and the reasons can be many..

    1. Still NagiosXI is a lot cheaper with its one time perpetual license than other big so called enterprise tools with their per device per month/year cost putting holes in your pocket until you die and leave this IT world.

    2. Nagios Inc. provides support for its open source and commercial versions on a per year renewal basis but its still cheaper than enterprise tools.

    3. RoI is best even if you get this product and hire a team of permanent staffs to manage it if you feel so, though it’s not too difficult.

    While Jonah might be talking more to clients but it all depends on how effective they are in utilizing this tool as he said there are happy and sad customers. I also experienced stigma of IT industry particularly CIOs, CTOs and IT directors going blindly for enterprise tools, simply because they feature in some gartner’s magic quadrant reports and their mentality that these tools are way matured and supported by their vendors and end up in spending a huge chunk of their IT budget spending simply because of their mental satisfaction that if the tool breaks it will be well supported by their respective vendors and thus falling prey to year-on-year recurring cost of these tools. While there are also scenarios where companies are seen using more than one monitoring tool to meet their requirements including the open source nagios core and in some cases using it as MoM (Master of Masters) because of it’s flexibility.

    In my experience I have done some migration projects where these enterprise tools got replaced by NagiosXI providing a good RoI to them. However I will say it all depends on CIO and so called IT decision maker’s choice that how confident and technical he is.

    While I have worked on core and latest nagiosXI versions I want to provide my honest points here..

    Pros and Cons of Nagioscore/XI

    Pros
    —–
    1. Highly scaleable, extendable and robust product both core and xi
    2. Supported by huge community, so there is no dearth of resources on internet about nagios.
    3. Creativity combined with skills can make it beat any present day commercial counter parts, I agree it will require effort and time.
    4. No dearth of trained and knowledgeable nagios man resources who are widely available and cost less compared to other commercial tools expert who quotes high because of their scarcity.
    5. Numerous addons that can extend its functionality simply.

    Cons (Positive kind attention of Nagios Ent.)
    ——
    1. Requires a more robust auto discovery feature in XI
    2. BPI/BSM component needs to be more evolved may be auto dependency mapping and dynamic mapping and managing the IT component relationship maps.
    3. Auto modelling and configuration of hosts and services out of box feature (though can be done using integration with chef/puppet) it can be achieved partly using check_mk using passive checks but not officially from nagios.
    4. Purely lacks ECE (event correlation engine) a must for any enterprise monitoring tool. Events filtering and alert formulation based on events.
    5. Out of box integration feature with CMDB a must have for enterprise tools.
    6. Nagios owned monitoring agents (though I like Nsclient and its best) as this will give confidence to customers.
    7. Nagios endorsed standard plugins for monitoring any IT asset or component as it’s a big confusion in the community which addon or plugin to use for a specific IT component. Something like they have done with wizards (But more nagios standard and supported official configuration wizards are required for monitoring many IT components, can’t simply rely searching on reliable plugins in nagios community)
    8. To be virtualization and cloud ready monitoring solution.
    9. Backup and Restore feature in XI Interface.
    10. Out of the box support for setting up Nagios Distributed setup.
    11. Agents push and auto install facility on monitored hosts.
    12. XI Interface to install new plugins/addons
    13. Better and more cusotmizable reporting portal out of the box instead of pnp4nagios performance reports with reports export feature.
    14. Interface to export the details of host and services configured

    Though many of these can be achieved using your own skills and effort and using community addons and plugins which are available already, it’s indeed a great and robust product.

    We do consultancy, implementation and support for many open source products including Nagios Core/XI

    http://www.clearinfra.com/
    http://manishkr.com/

    PS:- I don’t work for Nagios or is a reseller of Nagios

  • 56 youtube.com video emusic free trial   July 28, 2013 at 9:08 am

    Today, I went to the beachfront with my kids. I found a
    sea shell and gave it to my 4 year old daughter and said “You can hear the ocean if you put this to your ear.” She put the shell to her
    ear and screamed. There was a hermit crab inside and it
    pinched her ear. She never wants to go back!
    LoL I know this is completely off topic but I had to
    tell someone!

  • 57 Manish Kumar   July 30, 2013 at 12:53 pm

    Please read this too..

    http://blogs.gartner.com/jonah-kowall/2013/07/10/nagios-let-the-religious-wars-continue/

  • 58 W Sanders   August 1, 2013 at 6:37 pm

    I would expect any Gartner analyst to be a shill for the overpriced performance and monitoring bloatware vendors that underwrite their paychecks rather than advocate an open source, infinitely extensible product that nearly everyone uses.

    The main argument against Nagios is that it’s getting awfully old, with CGIs based on static tables, etc. There are other open source tools catching up with it, and maybe soon to surpass it.

  • 59 Jonah Kowall   August 1, 2013 at 6:40 pm

    W : We can write about whatever we see as being important (unlike other firms) actually I’ve written many notes on inexpensive and free monitoring, open source monitoring, and other areas. I would suggest reading up before you post defamatory and uneducated comments.

    Agree it’s old and difficult to manage, there are better open source products out there, but Nagios continues to dominate the open source monitoring space.

  • 60 W Sanders   August 1, 2013 at 7:09 pm

    Perhaps I overreacted to the lack of named alternatives of the post. A better title might have been “Got Nagios? It’s Not 1999 anymore.” I’ve done a few contracts, the other way around: the first task was usually replacing Big 4 “recommended” bloatware with Nagios (I once worked side by side with a Tivoli contractor for 6 months, he never got the product working.) These are places where there is some amount of scripting/programming skill in-house, which is necessary, because the Nagios plug-ins are just as long in the tooth as Nagios itself. Scaling? Taking a single instance to 500 or more hosts and 2000 or more services is no problem for a modern Xeon-class 1 or 2U server, the real problem is that the Nagios CGIs can’t display data at that scale. Scaling above that with multiple instances, using passive, agentless monitoring is not hard. I’ve merely trialed Zenoss and Zabbix, and next on my list are Opsview and Shinken.

  • 61 Jonah Kowall   August 1, 2013 at 8:11 pm

    Agreed with you on there, but the issue is that the UIs and configuration management are still lacking, even on the newer open source technologies. Additionally the plugin quality leaves much to be desired. I would also have a look at OP5 (considering they are a big contributor to Nagios v4) and Centerity.

  • 62 Martin Peacock   August 2, 2013 at 8:47 am

    A ‘discussion’ that can only come from the OSS community :-)

    I’d like to pick up a point made by Andreas:
    “Oh, and for comparison; Paying 2-20 dollars per device you want to monitor (what’s that for btw? per month? per year?) is quite a lot of money compared to what it would cost to hire a consultant to set up Nagios for you.”

    Which, of course, is undeniable, but that assumes the options are treated equally within the Enterprise – and they are not. In general, the larger the org, the less flexible the *internal* supply chain, which means it can be difficult to stray too far away from the traditional procurement model. In this part of NW Europe there are folk in Public Service entities that would be quite happy to try such solutions but while they are busy actually delivering PoC, the procurement arm sticks out a tender n OJEU & the opportunity is lost.

    That’s isn’t a purely EU thing, as well. I recall reading about a US hospital (LA I think) who wanted to invest in a more self-help strategy but found that the financial structures in place just didn’t support it. (I’ve tried/failed a number of times to go back & rediscover the source for that and likely will try again).

    That isn’t to say it is the case everywhere – but a tendency. Clearly there are exceptions as demonstrated by the success stories presented in this discussion.

  • 63 C. G.   August 3, 2013 at 12:09 am

    I’ve been using Nagios (and XI) for a number of years and love it. I’ve used it at all the companies I work at to great effect. I’m also a Nagios Veteran on the support forums and regularly help others with Nagios. If anything, Nagios users are LOYAL. (hence the comments)

    All that said, this article is mostly spot on correct. Nagios is extremely individualized in every environment. If I got struck by lightening, the company would be TOAST. Need some uptime reports for the uppers? Better brush up on MySQL because the out of the box reports are terrible.

    Cheap alternatives? I’d love to see some. I’d switch in a heartbeat. I need a solution I can show to executives and hand off to a junior admin to manage. Nagios is neither.

  • 64 Martin Phillpot   August 7, 2013 at 9:27 am

    Looking at all these positive responses I think Nagios is a very geeky toy and, if you know what you’re doing it can undoubtedly do great things, however, if you don’t, like me, it is a very unfriendly piece of kit that doesn’t do what the techs or the manager want. It seems to be very weakly documented and in some situations requires scripting to get it to do what you want it to do.

    What I need is a product that can go in to place quickly, alerts promptly, is simple to configure and can produce nice shiny reports for management at the touch of a button.

    I would dump nagios in a heartbeat if it were my choice.

  • 65 hdaz   August 8, 2013 at 4:56 pm

    @Martin Phillpot

    “What I need is a product that can go in to place quickly, alerts promptly, is simple to configure and can produce nice shiny reports for management at the touch of a button.”

    And what happens after you place such a product and wait the 1000 days 10,000 days or whatever this article claims nagios core becomes a mess by …

    You will find three things:

    One: the lovely shinny product does not do everything you want it too because you have xyz requirement that was never thought of at the birth of the product.

    Two: Its become a mess as well.

    Three: an emergency breaks out and you have no idea how to fix shinny monitoring tool.. The only place you can turn to is the tech support guys from xyz company reading from a script or some consulting firm which will come in and either tell you to use something else.. or charge you a fortune for pressing five buttons as they claim they had to contact marshions to get the answer(s).

    From your simple requirements you might as well spend your money on NagiosIX and press a few buttons. Because that way when you do want to take the monitoring seriouslly you have a great base to work from.

    @ The article / author
    As I have kind of outlined above its not always the product being used but the management and processes that are built that can make or break success.. okay nagios guru leaves the organisation where are the HR and management processes in place to make sure that the company is covered ?

    If everything was simple within companies whats the point of hiring any management layers at all?, forget the cost of employing tech and technical people, move back to a filing cabinets and a lock..

    The biggest problems are normally within management and the lack of ambition, lack of understanding and trying to balence not looking totally stupid or begging everyone to view ego as .

    APM Monitoring – surely the checks and plugins have been availible for as long as nagios/netsaint has been availilble… if your not so concerned about privacy use newrelic and others as well…

    Nagios Core has its issues no one is saying it is perfect but that uglyness is what makes is so useful and beautiful, thejava guy can come and write any checks he wants, the python guy can come along and re-impliment them or make others as they need and someone else can just use bash or whatever.. its all open its all availible and not hidden.

  • 66 Jonah Kowall   August 8, 2013 at 5:36 pm

    hdaz : Thanks for the comments, some very valid points there. I agree most systems deteriorate over time, and it’s strange none of the tool vendors try to prevent that issue from occurring. I guess the best case is if they make tools which upgrade regardless (such as a SaaS platform) then it should help keep the tools current.

    If you call Nagios APM then you should read up on what APM is, Nagios does infrastructure monitoring (which includes application instances, but not transaction level performance monitoring).

  • 67 svd   August 16, 2013 at 9:59 am

    That article makes no sense. It should be removed.

    I’ve tested and used several tools and nothing beats Nagios (+ plugins, like check_mk) on flexibility, robustness,notification logic,…. Extending Nagios with your own plugins is very easy! compared with other products.

    It’s true if you say Nagios core can’t do APM. But who uses Nagios Core exclusively. The true power lies in the thousands of addons and plugins. So in effect there are plugins that do APM and if not available for your particular application, you can write you own check for it. And if you really want you can buy a commercial APM and connect it to Nagios to have a central monitoring collection point.

  • 68 Jonah Kowall   August 16, 2013 at 11:55 am

    svd : there is no APM plugins for Nagios, please show me one which does transaction tracing. Nagios also is very poor at event correlation or building it’s own topology. It makes a inadequate single pane of glass. Everything is extremely manual, hence Nagios is very basic is the possible capabilities contained within the open source components. Most people I speak with feed Nagios data to other tools which can do these things.

  • 69 Henrik Blaaberg   August 21, 2013 at 2:10 pm

    I must agree with Jonah.
    Nagios does not do APM.
    Application Performace Monitoring is NOT monitoring how the application server is doing. A lot of the solutions based on Nagios that I have seen over the years call it APM where in fact all they are doing is monitoring the server resources and pinging from and external site. We at CapMon have worked with NetSaint and Nagios for the past 12 years and have build a lot of stuff around it to make it more usefull and manageable form our customers point of view. No doubt you can build probes or taps that can do what the big guys like Compuware are doing to provide APM. That’s just not there. We have developed a module (ResponseView) to provide lightweight APM, meaning looking at performance from an endusers perspective in the form of screen recognition and robot functionality. This is still not APM since we are not able to trace each transaction down to code level etc.. but it helps give an idea about what the endusers are experiencing on a selected number of applications and operations.
    Jonah’s comments from Clients are pretty much the same as we have seen and heard during our time with Nagios.
    I’m guessing that Jonah’s point here, (correct me if i’m wrong) is that you need far more competence to make a good installation based on Nagios, than what you need to get one of the simple programs that he mentions in his article to work reasonably.
    I also agree with most other comments in the thread, that you actually can build a solid and great monitoring solution based on Nagios, IF you have the knowledge and knowhow.
    Most companies that we have been in contact with over the years, are simply not willing to invest in having employees only working with Nagios.

  • 70 Chris   August 22, 2013 at 10:06 pm

    Johan is correct, Nagios sucks. You know what else sucks, Zabbix. Open Source doesn’t = gold. Especially when you have to put extra time and effort into the tool that is suppose to be saving you extra time and effort.

  • 71 Dave   August 23, 2013 at 6:23 am

    I was looking for information on a paper I am writing to get rid of Nagios in our multi client environment and found this post.

    We are in a situation of 30 people, and the one guy who knows Nagios left, we are now in a situation where we can’t customise.

    Despite what most of the people in this forum say, Nagios is a poor product.

    Highly customisable = Highly manual = job security for Linux admins

    I was looking at youtube videos of how to configure Nagios. After modifying the 50th cfg file I fell asleep. How prone to human error is it? it’s amazing how manual it all is.

    I think tools like Whats Up Gold are much much better. Simple to implement, all of the probes are out of the box, they work. Dependencies into ESX. Topology, it’s just all good.

    I am generally a SCOM guy, and I was surprised how hard Nagios is to use, how poor its coverage is (unless you get a couple of expensive *nix guys) and it’s overall crap experience.
    Nagios is severly limited in any modern environment when Business Systems level monitoring is required.

    I am with Johan, Nagios stinks, and I will be using his article and the Pro-Nagios comments (which are hilarious) as evidence we should dump it.

    The amount of places I have worked in where they used to have Nagios but they got rid of it is almost to many to list.

  • 72 Jonah kowall   August 23, 2013 at 3:27 pm

    Dave, I have seen it work, but I’ve also seen it fail for the reasons you mention. It really depends on the scale and what you are trying to do with it. I am in general agreement simple tools like Ipswitch WUG, Solarwinds, Zenoss, Zyrion, Sciencelogic, and other unified monitoring products do a munch better job at monitoring the network, but additionally can monitor systems and the base applications quite easily and effectively with much less pain. They also have hooks and APIs to pull data from common infrastructure products such as VMware, Hyper-v, and others with minimal work. There are costs involved, but they are minimal compared to some of the larger systems management players solutions out there.

  • 73 John   August 25, 2013 at 9:52 am

    I think most people are missing the point of the article here, Jonah is clearly talking about running Nagios in Windows shops, in non-technical companies.
    Deploying any technology that only one staff member understands and can maintain is a foolish idea, whether it be for monitoring or anything else.

    Most of the pro Nagios people on this thread are probably pro Nagios because they’re competent *nix admins, and it’s the right tool for monitoring *nix infrastructure. I’ve never met a competent *nix admin who’s not familiar with Nagios, but then again many of these same admins wouldn’t have a clue about some equivalent Windows technology.
    No person in their right mind would run vanilla core, without your own custom checks or try to maintain Nagios configs by hand.
    Everyone these days, at the least, uses some config building system, and generally some of the great plugins and addons. Even then a large proportion probably never ever touch the config files as their configuration management software of choice automatically does it all for them at build and installation time, automatically configuring dependencies on networks, vmhosts, etc… Infrastructure as code and all.
    And this is one of Nagios’ main strengths, no it doesn’t auto-configure, it easily integrates into your environment, and your environment configures it. It does what it’s told to, how it’s told to, and monitors what it’s told to, not a sub set of supported applications and components.

    So I have to agree with Jonah, if you’ve only got 1 “talented engineer”/”geeky linux guy” on your IT staff and not a team of skilled *nix admins and you’re trying to handcraft a Nagios config – you’re doing it wrong, ditch it, get a SaaS solution and stick to what you’re good at.

  • 74 Norman Harebottle III   August 29, 2013 at 1:37 pm

    I managed a small SaaS’s network operations center with a Nagios Core and really got to like using it. Yes, you have to edit text files to manage the configuration, but you CAN set those files up just about any way you want. :-)

    I currently use Nagios to monitor my application state and dependencies inside a large corporation because I can’t get anyone to give me access to the corporate monitoring solution! Oh, and its running on a Mac Mini (go Homebrew!).

    I’ve written simple plugins for Nagios Core in Python, C# (running on Mono querying SQL Server) and Bash shell scripts.

    I’m a happy customer. :-)

  • 75 Willy Dr. Takang   August 31, 2013 at 4:31 am

    Nagios in my opinion is a tool of circumstance i.e you use because you have to not because you like to. While the company I work for develops tools which are better than Nagios, they are too cheap to let their employees use these tools. I started using Nagios because I had a ton of equipment to keep an eye on and Nagios was the only free option I could readily lay hands on. The learning curve was steep, add to that the fact that we are running ubuntu servers and you have another issues to deal with since most of the documentation I found for it was written for Redhat and Centos. That said, it has been doing what I expect of it. There are tons of plugins and if there is something I need which does not exist I just go ahead and write a plugin for my own use.

    That said, while Nagios is working for me, I feel sad for the day when I have to leave the company. I am the only one who knows what I did to get Nagios working and the effort it took. So while the system is up and running I can only shudder at the thought of letting a software developer try to figure out what it is I did when I am gone.

    I hear a lot of complains about having to edit text files to use Nagios. In my opinion that is a plus. GUIs annoy me. GUI’s were made so that my grandma could use a computer. My opinion is when push comes to shove and you have no other options, use Nagios.

  • 76 Jonah Kowall   September 1, 2013 at 11:21 am

    Willy, very good points here. There are a lot of good free tools to handle 25 servers or so, but if you get beyond that you are limited with having no budget.

    While I do agree that a good API and text files are fine to have a good GUI helps those less experienced users use the product. It also helps with having mobile access, trying to use ssh on a smartphone is always painful, but doable.

    Thanks for the post.

  • 77 Panagiotis Stathopoulos   September 3, 2013 at 9:27 am

    We have nagios over some hundred devices, and approximately 1800 checks, from datacenter power, backup electrical generation operation, UPS status, to application server and db server status, to content sensitive checks. We get notifications from the dashboard to email or to SMS messages over a connected GSM phone.
    Anything that can be done with SNMP or an agent extending SNMP or a locally run script.
    The level of flexibility is astonishing and the cost extremely low.
    While its true you need some basically skilled engineers (any company with hundrends of server needs that) or that it can do anything out of the box (e.g. APM) the basic engineering behind nagios is solid.

  • 78 cheap article writing and submission   September 11, 2013 at 3:03 pm

    Inexpensive, Reasonably priced Expert Article marketing Services

  • 79 Chris   September 13, 2013 at 4:18 pm

    As a long time user of Nagios and NagiosXI I feel I have to put my two cents in. Some strong points you cannot beat with XI is the install, you run one script and you are basically done. The pricing cannot be touched by anyone. If you are strictly looking for monitoring then I am a firm believer Nagios is the way to go. If you are looking for Config management, software deployment and other management type utilities there are other products out there that specialize in this. I have yet to find a “perfect” all in one solution that is why we keep our NagiosXI system running side by side with our network management solution. You cannot beat NagiosXI support, their team for no extra charge helped us with code to customize a solution, you would be hard pressed to find that with any of the other solutions.

  • 80 Jonah Kowall   September 14, 2013 at 5:19 pm

    Yes Chris, it’s one of many Nagios clones, I see a lot more adoption of Centreon, OP5, and Centerity than Nagios XI. In fact the Nagios XI team doesn’t contribute much to Nagios compared to vendors like OP5.

    If you have to customize your monitoring solutions (which pretty much every Nagios user needs to do) then there is something wrong. Chris, which network management solution do you use?

    There are ways to make simple tools that handle network and system monitoring work, products like ManageEngine, Solarwinds, and even Ipswitch can handle both in a simple and straightforward manner for basic metric collection. More complex needs should be handled by APM tools which go deeper into the application systems.

  • 81 Srinivas   September 17, 2013 at 8:19 pm

    You are right.

    This is lil challenging integration need enthu IT guys to take over. Sys/Network Admin is a skilled guy, He can easily understand this.

    We are currently Using Nagios / Nagvis / { Perl Scripts post Free IP List } / Integrated KWOK to the dashboard.

    Did you done your HomeWork before you write this ?

  • 82 RonJon   September 20, 2013 at 8:45 pm

    You say: “The time and effort needed to manage this software is much better spent buying a simple monitoring tool to get the basics covered for infrastructure health.”

    But that is the point. 99% of all alerting is based on values, thresholds, up/down conditions. In Nagios Core if I can get a value from say a JMX query, the log scrape of some shop floor machine, or the freaking coffee maker, I can alert on it.

    Name 3 other monitoring products that offer their customers that kind of customization? Please include the cost per node for those other, user customizable, products.

    I just interviewed at Apple last year, they even use Nagios – and they could buy anything out there!

    We have a couple other open source products for trending and network analysis, but lately I find myself in the pleasant position of asking other departments, “What else would you like monitored?”. In most every case I use a Nagios plugin or simple script to monitor what they want.

  • 83 Victoria   September 22, 2013 at 10:00 pm

    Thx!, gran aporte, ya te contaré!.

  • 84 Gin   September 23, 2013 at 11:18 am

    It is clear that research is not done properly because “your t-shirts does not fit on everyone’s chest”

  • 85 Marcelino Cowper   September 26, 2013 at 2:50 pm

    Surprisingly revealing many thanks, I do think your readers may well want further blog posts such as this carry on the terrific perform.

  • 86 Ian Fraser   September 27, 2013 at 11:20 am

    So are you saying this in regards to Windows, or all platforms (Linux, Unix, Apps?)

    Yes you can buy other tools but that’s not the point. We use SCOM for our windows estate because of its in depth management set for low level windows monitoring, and its been tailored by Microsoft for Microsoft platforms. I think most organisations would agree that with the diverse set of windows platforms and infrastructure apps like exchange, AD SCOM is the best bet. HOWEVER, it doesn’t discount Nagios from monitoring other areas of the estate including apps.

    The architecture can handle feeds from any app out there, including SCOM if you want to tailor it (and any decent engineer can do this). Then why not feed critical alters into Nagios and have a high level view of the entire estate?

    Your comment on an engineer leaving the company in the lurch. Have they ever heard of documentation??!! You document your build, settings, operations guides, high and low level design, build docs.

    It sounds like someone’s told you its rubbish and you’ve never actually verified the source.

    Poor from both a journalistic and technical perspective I’m afraid.

  • 87 Jonah Kowall   September 27, 2013 at 2:07 pm

    Ian, the monitoring Microsoft does with SCOM isn’t different than what most other monitoring tools do on windows, they don’t collect any additional detail aside from what’s already available in WMI and the event log. It’s all standard stuff. With SCOM you can only have 1 version installed, Nagios on the other hand is a conglomeration of packages which work together. This makes documentation and upgrading difficult when the person with that knowledge leaves. It happen regularly on my calls.

    In terms of feeding data to Nagios, the visualization layers and event correlation capabilities in nagios are next to zero, Microsoft is much better at both use cases, but still rather weak.

    I’ve used all of these tools in production environments, I was a practitioner and manager at several startups as well as large companies. I also continue to test products in my lab to stay current aside from what the vendors and end users tell us. I will speak to over 800 (shooting for 1,000) end user clients this year about monitoring technologies on my calls.

  • 88 Arie   September 27, 2013 at 6:00 pm

    Nice Article that can help people to make choices, but not complete ihmo.

    I have recently installed a monitoring system based on Nagios with Check_mk (+400 nodes/+6000 services). This combination is very powerful. Configuration of the monitoring get’s very easy, and can be done by hand in just a few configuration files or with the Multisite interface witch is part of check_mk. Anyone at our site with little understanding is capable to use and configure this monitoring system.

    A very powerful thing that is in progress is the OMDdistro. Just install Nagios and everything else needed by one rpm package and your done. Configure multiple instances of a monitoring server on the same node (prod and test) just with one command. And don’t forget the extreme powerful monitoring capabilities of Nagios 4.

    Basically I did not like to start with Nagios core, it is like a child without arms and legs, but with some add-ons it became a nice and friendly kid in the and. And what is wrong with that !!

  • 89 Rob Hough   September 30, 2013 at 4:21 pm

    “provide much greater business value to the non-technical users.”

    I am not certain why an administrators tool box needs to provide greater business value to non-technical users???

  • 90 Jonah Kowall   September 30, 2013 at 4:51 pm

    That’s because you are thinking tactically when monitoring versus strategically. Investments in infrastructure are approved and vetted by non-technical business users, hence you must show how things are running. If you take things further the monitoring tools can even track business metrics, something commonly done with APM products (business metric extraction and transaction monitoring). There are many use cases and needs for these tools to provide non-technical visibility.

  • 91 Rob Hough   September 30, 2013 at 6:07 pm

    I get that executives need metrics to help guide a successful business model. It just sounds like many of them forget that part of that success is reliant upon a stable, working environment. What tools the administrators use to accomplish that goal, within their budgets, should really be irrelevant to that equation.

    Have any of these APM tools been named, as solid replacements for nagios?

  • 92 Rob Hough   September 30, 2013 at 6:44 pm

    P.S… I am asking because we are currently lacking these tools. I know nagios, but I am not against looking into other tools as well.

  • 93 Jonah Kowall   September 30, 2013 at 9:56 pm

    Rob : Nagios handles server monitoring and basic network monitoring, APM tools focus on the application layer versus the infrastructure. We see a lot of the unified monitoring products doing better in terms of looking across the infrastructure and incorporating views from the application perspective, but APM should be a better investment generally speaking. The commercial Nagios products help with better manageability and visualization, while still having the open community and architecture many people prefer (as you can tell by this massive thread).

  • 94 Johnny   October 1, 2013 at 8:12 am

    Jonah: I dont know much about monitoring but this just caught my eye. A comment of yours in the thread.

    “I’m not saying Nagios is bad, I’m saying the open source tool ends up costing significantly more than the license price for most environments. You can read countless research articles talking about the cost of open source, sometimes it’s a good bet, and other times it is not. Nagios is one example where the open source isn’t a good bet. ”

    I would like to point out that “Open Source” has nothing to do with a product being enterprise / commercial. Op5 as you mention “sells” an “Open Source” version of Nagios.

    I bet there’s alot of research on the cost of open source, but alot of people get atleast this wrong. Commercial software can and is open source.

  • 95 Bernard Hoffman   October 7, 2013 at 9:17 pm

    I’ll have to agree AND disagree. With 24 years as both an IT insider and consultant, there are many tools that are easier to use and integrate into your IT fabric. HOWEVER, many of the commercial tools mentioned in this thread are orders of magnitude more expensive to acquire (not to mention annual maintenance).

    I would be interested in a TCO analysis, but would still bet that Nagios (in either the Core or XI platform) is more cost-effective than competitors. To add, I’ve never seen a product that does APM well without highly specialized add-ons that make the OVERALL infrastructure performance reporting more expensive to produce.

    I’d love it if Nagios were easier to deploy, manage, upgrade, and replicate in the enterprise environment; and it seems that the Nagios team is working towards tighter integration than I’ve ever seen (considering I started with NetSaint…).

  • 96 Lane Williams   October 9, 2013 at 6:55 pm

    I’ve been using Nagios since the NetSaint days. Yes Nagios does have a lot of complexity. That complexity equates into interoperability, flexibility, and usability. How many ‘itys’ can I use?

    Nagios monitors anything and graphs and reports everything. I use both Core and XI, and it satisfies everyones needs from the admins to the CIO. We monitor network flow, system stats, application stats and services (JVM’s, web, databases, etc.). Business process monitoring. We monitor and report on Netezza and Greenplum analytics. We monitor data center environmentals/power and generate historical reports. We generate dynamic reports from Nagios data…I can go on and on.

    I guess if you are new to Nagios you just don’t know what you can do with it and what is out there. But that goes with any package. You ever have to setup a Tivoli product or HP Openview or NetIQ? They all require effort and expertise, and you have to script with all of them if you want to monitor your entire environment. Simple monitoring tools are just that, simple. They do not offer the depth that an enterprise data center needs to keep the lights on.

    Our entire team understands Nagios, as our entire team understands all the applications we deploy. Within IT, to have just one person on a team understand a product is just bad management.

    If you are strictly a Microsoft shop with no Linux/Unix admins, then Nagios is probably not for you. Not that a Microsoft admin can’t install and manage Nagios, Linux might be a little of a learning curve.

  • 97 Free Infrastructure Monitoring Tools : Middleware tweaks   October 10, 2013 at 6:53 am

    [...] However there are some people out there who would advise caution against using Nagios like this Gartner Blog. As you can see from comments on that blog that there are many people opposing views presented in [...]

  • 98 David Scholefield   October 16, 2013 at 8:23 am

    This article has the whiff of the corporate consultant selling the overpriced ‘nobody was ever fired for buying IBM’ about it. It seems to be supporting the general attack on open source, highly skilled approach to big business I.T. where companies like Gartner make most money from funnelling clients into an expensive, standardised, route rather than the (often equally as functional if not better) free route where there is less money for consultants like Gartner.
    I’m responsible for managing the monitoring of high availability networks mostly based on LAMP stacks and Nagios is easy to implement and manage, and it works extremely well. This kind of article is just misleading fluff.

  • 99 Jonah Kowall   October 16, 2013 at 1:58 pm

    David, I’ve replaced a lot of IBM technologies in my day, and have been an open source contributor to FreeBSD and other projects, so that couldn’t be further from the truth. The fact is that not everything open source is of good quality or usable.

  • 100 Miguel DeCamps   October 17, 2013 at 7:31 pm

    Jonah you really need to do more research or possibly consider an alternate field.

  • 101 Jonah Kowall   October 17, 2013 at 7:39 pm

    Miguel, that’s not what your colleagues say who’ve worked with me in the past :) I’ve worked extensively with dozens of monitoring technologies out there, and hence my research reflects my experiences as well as speaking with clients who implement and manage these technologies.

    I appreciate your opinion and non-constructive criticism.

  • 102 Miguel DeCamps   October 18, 2013 at 2:53 pm

    Well alright you have a point it was not constructive of me to write that; but the “Got Nagios? get rid of it” title is a bit hurtful I felt the jab :-)

    I will agree with you that to stand-up a Nagios infrastructure is not a simple task and by all means not for the faint of heart. Staging a Nagios infrastructure is complicated and takes a wide array of different skill-sets to properly deploy it, but if you do have talented engineers and possibly sell managed network services as a core business product I do believe that there is a lot of flexibility a solution built on top of Nagios can bring to the table that may be well worth looking into.

    I would not have gone as far to say “Get rid of it” (It is catchy though!), but if your gig is not Monitoring and you want something much simpler to deploy by all means as you mention there are a myriad of off the shelf solutions out there that companies can utilize for simple health checks of their infrastructure.

  • 103 The Consensus Is...   October 22, 2013 at 6:32 pm

    I’m more inclined to follow the consensus of advice I find across the entire internet on multiple boards, forums, mailing lists, and blogs than the rants of one sysadmin. Hmmmm.

  • 104 RonJon   October 31, 2013 at 7:19 pm

    Johah,

    How come no answer joy for my previous comment? Please respond.

    Sure maybe you run this in your lab and maybe plenty of other monitoring products. But for those of us in the real world it’s about the customization man! At my medium-sized company we don’t have the money for a Gartner touted product, but I looked at quite a few of them before settling on Nagios (for a second time). It’s actually Fully Automated Nagios, but Nagios is the alerting engine. I have found nothing that offers me the extreme flexibility and customization of Nagios. Proprietary products might say they can do anything Nagios can, but you had better be willing to wait and it will usually cost you $$$.

  • 105 Marc Mazerolle   October 31, 2013 at 8:59 pm

    What do you expect ? Would Gartner ever recommend an Open Source software ? Could they ?

    Follow the money trail. Who generates revenues for Gartner ? Hardware and software companies like Microsoft, IBM, HP that all have paid big money to have their products in a “magic quadran”. And they all have products competing with Nagios…

    I will make my opinion based on impartial sources.

  • 106 Richard   November 2, 2013 at 8:16 am

    There are some good discussions/arguments/opinions/facts in these comments. Thanks to those who posted. The information within the comments were by far more valuable than the contents of the article.

    From the comments in this article– it’s clear that different monitoring tools are more suitable than others depending on the needs of an organisation. Use the tool that is as simple or complex as you need it to be. Figure out what value the product will bring to the organization, and if the cost fits the budget. I mean, what do you want and what do you need? When will the cost of the product outweigh the cost of your time setting setting it up?

    For the author to be someone who apparently works in the industry, you would think he could appreciate ‘best tool for the job’. He seems like a hotshot with an inflated sense of self-worth. Typical.

  • 107 Jonah Kowall   November 2, 2013 at 1:00 pm

    Richard, the fact is that I see more failures with Nagios than with other tools. I speak to over 100 Gartner clients per year using Nagios.

    There are much better open source alternatives out there which have the same flexibility without some of the legacy baggage. Most Nagios users don’t particularly like the product and usability, but it’s the de-facto monitoring standard for no good reason.

  • 108 Zachariah   November 5, 2013 at 6:57 pm

    You speak to over 100 clients a year about Nagios, but you need a “Briefing” to figure out how it works and what it can do? Real nerds who talk to other real nerds, like Nagios. Nerds are why it’s the de-facto monitoring standard, because when your briefings fail as guidance to management, they ask us what to do.

  • 109 Jonah Kowall   November 10, 2013 at 5:27 pm

    There is more to a solution than your own testing, hearing it from others, and your past experiences. There is more to a company or product than technology. Wait for my upcoming post about some issues which have been causing issues within Nagios.

  • 110 j-marr   November 11, 2013 at 9:41 pm

    Real, actual companies also use nagios without ‘failures'; successful companies with smart, internet-savvy people making decisions – maybe this is too anecdotal but this is a real person at a real company and you can’t argue with the results:
    http://37signals.com/svn/posts/3178-nagios-monitoring-performance

    FTW: nagios + check_mk (w/livestatus) + configuration manager ie puppet/ansible/cfengine/etc

    I think it addresses all negatives mentioned except maybe needing someone in IT to manage/configure/build it but it’s an IT tool so I’m not understanding why that would be a negative…

    The 37Signals article inspired me to set up a similar infrastructure at my company and it is blowing people away. The amount of data collected is staggering, it can scale for enterprise and beyond and we have noob users logging in and managing their stuff. if you want to get close to the metal then you need to learn python – whoa, another IT related skill… put it on your resume and a winner is you: http://lmgtfy.com/?q=win+win

    Remember, setting this env up and learning it is time spent upfront when you have it and it saves you time over the long haul or when you don’t have time ie putting out urgent fires.

    …or you could pay a lot of money for a boxed solution and someone to call on the phone lol

  • 111 Jonah Kowall   November 11, 2013 at 11:19 pm

    Lots of the products on the market are low cost, especially for simple things like server monitoring. Nagios focuses on a single aspect of most infrastructures (the servers), but does a pretty poor job at monitoring storage, virtualization, and the network where topologies are more critical. I do agree you can put together good solutions with open source and I speak to clients doing this regularly. More advanced tools such as statsd, collectd, graphite and other associated components contain less state and are significantly more modern than Nagios. The issue is that those tools require a lot of work to integrate and maintain. I’ll have upcoming posts on the major changes happening with the Nagios open source project in the next week or two.

  • 112 j-marr   November 11, 2013 at 11:39 pm

    I guess the problem I’m having here is that a theory only remains so as long as no one single fact discounts it.

    the fact that at least 1 successful company uses Nagios+ successfully blows out the Nagios-is-no-good theory.

  • 113 j-marr   November 11, 2013 at 11:47 pm

    sorry I should mention that I appreciate the article and the discussions it inspired regardless of which details I agree or disagree with. I’ve not read your articles before but I look forward to doing so from now on.

    thanks Jonah

  • 114 Ademola   November 12, 2013 at 2:36 pm

    “The fact is that not everything open source is of good quality or usable.”

    It wouldn’t be nice to place Nagios in this category. Nagios is a good product and the more you use it the more you understand how powerful it is. If you are looking for a click-a-link product, Nagios is surely not fully there yet but it gives you power to get want. The products focus had remained for relatively many years a monitoring tool that is quite flexible to monitor virtually anything.

    Now pause and think, which one of those pay-a-tonnes click-a-link monitoring allows you to “monitor virtually anything”? Ok pause again and think about – “virtually anything”?

  • 115 Aaron Toponce   November 12, 2013 at 7:32 pm

    Nagios is the de facto standard for go-to monitoring, and there’s good reason: it’s so flexible, scales exceptionally well, and is highly configurable. And it doesn’t take much to get caught up on what it mostly offers.

    I’ve tried many monitoring solutions- Zabbix, OpenNMS, BigBrother, and others. None of them offer the clean configuration and flexibility that Nagios offers. For example, maybe I want to tie in IRC and XMPP push notifications. With Nagios, it’s a piece of cake. With others, not so much.

    It’s also trivially easy to write Nagios plugins. Anything from simple shell scripts to complex Python or Ruby applications. And because it doesn’t require a backend database server, but rather is just plain text, Nagios integrates very well with configuration management systems like Salt.

    The best part though, is the learning curve. I’ve been a UNIX/Linux systems administrator now for 15 years, and I’ve seen plenty of “newbie” admins pick up Nagios in a single 8 hour shift, with no previous experience, and have it monitoring a handful of services and hosts.

    In a previous comment, you said ” [Nagios] does a pretty poor job at monitoring storage, virtualization, and the network where topologies are more critical.”

    Curious what you have to back that up, because we’re using it to monitor BackupPC, and alert on old or failed backups. I have it in my personal setup monitoring my GlusterFS 2-node storage cluster, and also monitoring my ZFS RAID array. I have it monitoring several virtual machines, and it’s trivially simple to monitor Cisco and Juniper routers/switches with SNMP communities.

    I’ve personally either seen or administered Nagios installations monitoring particulate matter in the air, cooling and heating systems, humidity, door locks, biometric scanners, fan speeds, and even weather patterns. If you can get data into the computer, you can monitor it with Nagios plugins.

  • 116 Jonah Kowall   November 12, 2013 at 7:36 pm

    Aaron : when you deal with larger environments there are several topologies which need to be discovered and monitored to avoid event overload and allow for supression:

    Storage -> Server -> Hypervisor -> Guest OS -> Network (multiple devices interconnected)

    If you don’t have the linkages and topologies of the tiers it becomes more complex to monitor and troubleshoot. This is something which even basic tools can handle, but Nagios cannot handle topologies which are discovered, the administrator must manually build the hierarchies and dependencies.

  • 117 Aaron Toponce   November 12, 2013 at 7:43 pm

    Just to show you what you can do with Nagios, check out this site that is monitoring:

    * Volcano status
    * Earthquake levels
    * Storm alerts
    * Particulate pollution
    * Avalanches
    * Number of buses running
    * Airplanes in Iceland airspace
    * Banks
    * Exchange rates
    * Surgeries
    * People in ER
    * …

    Let me see that with some other solutions, open source, proprietary, or whatever.

  • 118 Olivier Contant   November 15, 2013 at 7:02 pm

    Hi Jonah,

    I’ve read many of the comments but I stopped at the middle as it was really long. So sorry if I repeat something already said.

    I kind of understand where you are coming from and why you wrote this article. I agree that Nagios is not userfriendly for non technical Unix people. As such, any tools related to Unix is not advisable for them. They should stick to Windows solution and overprice solution that leverage the technical part.

    But to completely avoid Nagios, I must admit that I cannot agree with you. At the moment that you have thousand of host and as much services, you need something that can scale, can be distributed and HA. Commercial product that can do that are way too expensive for most business.

    You say that Nagios is limited.
    I find Nagios very flexible and as any flexible tools, it has his drawback of being complicated. More control = More expertise. Even so, the tools do what WE want.

    For the drawback of complexity to maintain, as system admin, we have a collection of tools in our toolbox to patch thing up and make it easy to maintain. Windows, you have one utility that do thing in a limited way. Unix, we use a collection of small very reusable tools to do the same thing but with way more flexibility. This is our philosophy, we patch many small tools altogether to do something bigger.

    In this case, *Nagios, Puppet, and so on… * So again, yes for a junior or non technical person, Nagios can become a bit hard to maintain just like any Unix system would be for any beginner.
    It’s like in anything, you don’t give a Ferrari car to a 16 years old kid that just got its license, he will crash it. *Probably a Justin Bieber influence on this analogy*

    Something that most new comer will not take the time to check when they configure their Nagios system:
    http://nagios.sourceforge.net/docs/3_0/tuning.html

    Have you seen some of the project to allow a distributed and scalable monitoring architecture with Nagios?

    Project like:
    Icinga 2 (Fork of Nagios),
    Shinken (Fork in Python),
    Multisite,
    NRDP,
    check_mk
    Mod-Gearman
    Merlin,

    A bit more fancy:
    Krolyk, is a daemon which consumes Nagios check results from RabbitMQ and writes these to the Nagios command pipe. http://www.smetj.net/wiki/Krolyk,

    Moncli is a generic MONitoring CLIent which executes and processes requests on an external system in order to interact with the host’s local information sources which are normally not available over the network. http://www.smetj.net/wiki/Moncli

    Aggregating Nagios Checks With mcollective

    Sensu – Uses RabbitMQ and Pub/Sub Redis to scale the checks collection https://github.com/sonian/sensu

    There is new interesting project I haven`t yet to test like:
    Reconnoiter

    And for more granular and graphical data analysis, you can use OpenTSDB to generate specifics graph or graph on request and create alert to send to Nagios when data is outside the limit. OpenTSDB which is more Developer side tools, could be use to create specific graph for upper management. You just need to feed the engine with data you want to graph.

    What about you reply with a list of monitoring tools that you believe operate better than Nagios while offering the same level of Flexibility, Capacity, Scalability, Distributivity and High Availability.

    I would be curious to know about such solution and more particularly if they are more adminfriendly / userfriendly. If it exists, I believe it would not be possible to buy. It would be so expensive for business like Twitter, Facebook, Google, Yahoo with multiples 10 000 servers sites or like ours with above 3000 servers in multi geolocation site.

    Wish you the best,

    Olivier
    Informative source review of Nagios
    http://jedi.be/blog/2012/01/03/monitoring-wonderland-nagios-the-mighty-beast/
    http://www.slideshare.net/nagiosinc/john-sellens-nonobviousnagios

    Page of the differents projects list above for your review.
    https://www.icinga.org/icinga2/
    http://www.shinken-monitoring.org/
    http://www.thruk.org/
    http://mathias-kettner.de/checkmk_multisite.html
    http://mod-gearman.org/
    http://exchange.nagios.org/directory/Addons/Passive-Checks/NRDP–2D-Nagios-Remote-Data-Processor/details
    http://www.op5.org/community/plugin-inventory/op5-projects/merlin

    http://opentsdb.net/nagios.html
    https://labs.omniti.com/labs/reconnoiter

    http://www.devco.net/archives/2010/07/03/aggregating_nagios_checks_with_mcollective.php

  • 119 ครีมหน้าขาว   November 17, 2013 at 10:32 am

    Just want to say your article is as astonishing. The clarity in your publish is just spectacular and that i could think you’re knowledgeable on this subject.
    Fine along with your permission let me to grasp your RSS feed to stay up to date with
    imminent post. Thanks a million and please carry on the
    enjoyable work.

  • 120 Rms-Mit   November 17, 2013 at 11:17 pm

    Use Check_mk. It greatly simplifies and pulls in most basic checks by default. Easy to install especially if you use the OpenMonitoringDistro

    see talk
    https://www.youtube.com/watch?v=0KewneiTgJk

  • 121 Pablo   November 20, 2013 at 4:19 am

    Which “simple and inexpensive” monitoring tools are you talking about for $2-20 per device?

    Usability and sophistication are zero? Sounds like you didn’t even use it.

    As for ugly, who needs pretty HTML5 graphs and interface when told your server is down?

    I setup Nagios from scratch in one day, after never using it before. Its already monitoring 8 of our servers, and will tell me the health of Exchange Queues, DAG’s, and others, all for zero dollars.

  • 122 Jacques   November 21, 2013 at 7:30 am

    I have been using Nagios for about 3 months now to monitor our infrastructure. I did not know Linux or much about nix* for that matter before I started.
    On my first attempt I found that the it was not difficult to install the product on RHEL. It was further not that complicated to get basic monitoring configured but the fact that the free product to not have any configuration or auto discovery tools is a BIG problem.
    I also could not find updated plugins to monitor our IBM products, MQ, WAS, BAM, Process server etc. Most of the plugins were outdated and no longer supported by the author and community.

    So I started to develop my own plugins. Not very cost effective.

    I then decided I want to configure the full path to the servers including all parent child hierachies. What a pain. Nagios dop not have the ability to perform auto discovery of the paths to servers. You need to create this paths yourself and this would then be static paths. When changes happens in the network you would have to reconfigure Nagios. Not very effecient.

    Allthough the tool is free, my time to configure Nagios and develop plugins and make the product usable by far outways the advantages we get from the product.

    If you are penny wise and pound foolish then get Nagios. You will save allot of capex on your budget but opex will go up drastically.

    Ps. I have looked at a product call Managed engine which are miles ahead of Nagios. I would compare Nagios to Managed engine like comparing DOS to Windows 7 (not windows 8 , thats crap)

    I am sure the people that like Nagios are just protecting their own investment in the tool and even some job protection.

  • 123 Tom   November 26, 2013 at 3:30 pm

    Its the title drew me into reading this blog.

    The amount of content and reasons the author used in the article to persuade users to ditch Nagios is astonishing!

    That remind me, when I read a research report from one of our junior engineers year ago, titled ‘Linux vs Windows’. The junior engineer just put down few rubbish paragraphs, and concluded, Linux is free and open source, so Linux is miles better than Windows. So he recommended the company to develop their future applications on Linux instead of Windows.

    That was the worst report I have ever read, and this article resembling it!

  • 124 Jonah Kowall   November 26, 2013 at 3:32 pm

    Tom, this is a blog post it’s not a research document. This is based on client discussions, please don’t take the blog as the same depth and level of research content published.

  • 125 Emergency Exchange Support   December 10, 2013 at 12:35 am

    Hi the information on this blog is just amazing it keeps me coming back time and time again ,personally i met my wife using this site so i couldnt like it any more i have done my best to promote this blog as i know that others need to read this thing ,Thanks for all your effort spent in making this fabulous resource ! ok,nice one Jake

  • 126 Venugopal Maddukuri   December 11, 2013 at 5:16 pm

    Hello Jonah

    10 months later and this topic still draws responses. Its title is ‘sensational’.

    I often advice clients, tools come last in the decision cycle. It does not matter which tool you deploy, if you don’t get the business requirements and design right.

    I have seen customers with IBM Tivoli / BMC / Cisco / Microsoft / open-source (…name your product) installed and not used 5 years later, as nobody has the skill to understand the needs of business and translate into a relevant monitoring strategy/requirement.

    Implementing ‘inexpensive’ or ‘expensive’ tools does not guarantee great services.

    People who pay attention to the details and make things work – do guarantee results.

    Cheers
    venu

    Disclaimer: I used Nagios and loved it. It was complex too – forced me to understand my network and application integration requirements.

  • 127 Dale Lancaster   January 13, 2014 at 12:42 am

    We live in a modern GUI, “instant gratification” age. The author seems to believe that if you can’t click a button and have it automagically work, something must be wrong.

    Ultimately you have to have someone sit down and learn this stuff – anything really – that has any significant functionality. Even Microsoft sees the value of non-GUI interfaces – aka PowerShell.

    Nagios is extremely powerful, but yes it does take some old fashioned work and someone with slightly above average IQ to tackle it and deploy it properly. We have compared to other commercial products and they basically don’t match the flexibility and extensibility of nagios.

  • 128 Joe Brouillette   January 15, 2014 at 8:56 pm

    Hi Jonah, I must say you’ve continued to show a high level of professionalism in spite of some inflammatory/accusatory posts.

    I work for a small non-profit (12 servers & about 60 workstations) and I can’t afford some of the upscale APM and NPM programs out there so I’m running Nagios Core on a FreeBSD server and it works very well for me. The one comment you made about someone else having to come in and take over really gave me pause. Maybe I should consider a more mainstream program.

    You’ve mentioned that there are some very reasonable and effective products out there and I was hoping you could list some that might work in my shop.

    Thanks,

    Joe B

  • 129 Joachim Hoernle   January 18, 2014 at 5:47 pm

    Haven’t we all seen these heated discussions and hostile posts such as “what is the superior operating system?” “42 reasons why Android sucks.” “There is no programming language but Small Talk.” and so on. I need to say, that I’m fed up with all that useless debates. Why not combine the best of various worlds? Ok, all tools in IT operations have their pros and cons. Nagios was not made for the enterprise. Some commercial tools do you scale down, are expensive and integration doesn’t come for free. But what is the alternative? Ideologic discussions do not cause any progress, do not help to move forward any real implementation. Even worse they narrow the perspective of people so that real opportunities are overlooked. “Fragmentations works. In theory. Integration works. In praxis.” (anonymous)

  • 130 Ethan   January 20, 2014 at 5:32 pm

    Got Nagios? Get rid of it.

    I do like that title, but is there any option? I think we will see next generation of Nagios under a total different name. Nagios INC (Ethan) is closing down Nagios as a project step by step and kicking out all friends, one by one! They don´t allow people to talk free about the project. There is a reason why the community moving to next generation of the project called Naemon.

    So, back to the subject! Got Nagios? Get rid of it, and get Naemon or any of the real open source project out there.

  • 131 Jonah Kowall   January 23, 2014 at 6:41 pm

    Paying close attention to Naemon, good luck!

  • 132 OpenSource forever   January 27, 2014 at 3:21 pm

    Nagios is not the only way to solve the monitoring problem without no budget. Because one recurring plot here is “for free”, as free beer, gratis, zero cost. It’s incredible how nagios people ignore other amazing opensource solutions, growing and evolving, including new features and adapting to the new technologies: Zabbix, Pandora FMS, and don’t forget the “classics”: OpenNMS, Cacti, Ximon and of course, my favorite, ganglia.

    Please, Nagios, don’t consider yourself as the only one.

    Thanks for keeping posting Jonah ;)

  • 133 Ron   January 27, 2014 at 10:45 pm

    Joe B,

    Check out LogicMonitor.

    Good luck!

  • 134 accommodation rovinj   February 10, 2014 at 7:34 am

    elpful information. Fortunate me I found your web site by accident, and I am stunned why this coincidence didn’t took place in advance! I bookmarked it.

  • 135 Jason   February 13, 2014 at 11:47 pm

    Everything you say is bad about Nagios is just as bad as any other solution. The fact that you need to be familiar with more technical knowledge to manage Nagios just means that you will more easily be able to optimize and administer other aspects of the system. My company pays for one of those licensed solutions, and they are just as bad, not to mention they are not as customizeable and you end up sacrificing more efficient resource usage to pay for something that is slightly more dummy proof? In my opinion if you don’t have anyone technical enough to understand how to set up a basic Nagios system on your staff, your IT staff is terrible in the first place and you might as well just outsource everything. I don’t feel that you really provide any good points about why NOT to use Nagios that don’t apply to any other monitoring solution.

  • 136 Jonah Kowall   February 14, 2014 at 6:02 pm

    I would say there are many solutions which are as bad if not worse than Nagios, and while I agree that Nagios requires more self-reliance on your sysadmins the problem is that the solutions are implemented completely inconsistently making ongoing maintenance and management a problem. Customization is great when you need it, but you often pay for doing that down the line. Monitoring should be easy, there shouldn’t be a need to customize availability monitoring, if you customize it you are most likely making the tool do something it’s not designed to do. So finally, the problem isn’t the implementation, but how it looks 3+ years later when people have moved on…

  • 137 Rob   February 18, 2014 at 7:06 pm

    This article is so wrong on so many levels, it is hard to know where to start.

    Friend, you don’t know enough about this subject matter to be blogging it. Please stop, for your own good.

  • 138 Jonah Kowall   February 19, 2014 at 1:41 am

    Very astute commentary Rob, appreciate your feedback here. having implemented and managed Nagios, and also hearing from dozens of users about these issues give me a better fact base to make these comments. Glad you enjoyed reading it.

  • 139 Anonymous   February 19, 2014 at 11:25 pm

    The title of this article is terrible.

    The problem with everyone’s opinions about monitoring tools is that they think there is a solution for all scenarios. They figure that if tool X has feature 1 2 and 3 and tool Y only has feature 3 then somehow tool X is the only logical choice.

    What if you don’t need feature 1 and 2 and the inclusion of those features dramatically complicates the architecture and operational management of the system?

    Why the hell do I need to install a RDBMS to do some pings and port checks? Why do I need agents installed on all my boxes to monitor them?

    If all you need to do is ping a server, check a port, or write a completely custom plugin to do incredibly complex things and report back a status, then why use anything other than nagios?

    I’ve been using nagios since it was netsaint. I’ve run at least one nagios instance in every job I’ve ever had since 2001. Why? Because it just works and it has all the flexibility I need for writing custom checks or notifications.

    Someone in our organization recently decided that we would move everything to zabbix for monitoring. It’s been months and the system is still incapable of functioning reliably or in parity with the existing nagios installations.

    To people who use this blog article as a information source to decide what to use, figure out if the tool meets all of your needs while keeping your solution as simple as possible. Download everything, test it, play with it, break it, etc. Only then can you make an educated decision on which tool or combination of tools to use.

    If you’re concerned that your sysadmin might install some tools that you can’t manage in the future, perhaps you should better manage your sysadmins to document their work. That’s what documentation is for.

    Finally, if you want to fundamentally change the way you think about monitoring, take a look at streaming application data through systems like riemann (http://riemann.io/index.html)

  • 140 Jonah Kowall   February 20, 2014 at 6:37 pm

    Not sure why using a DB is a sticking point, considering most storage even on small devices such as Android phones uses relational data storage. Being able to store, recall, report on, and use the monitoring metrics and data is critical. Nagios doesn’t do this well, which is yet another knock against it and it’s dated architecture.

    If you need to check and end to end process or job you are probably using the wrong tool by writing complex custom checks in Nagios.

    Agree with you on Zabbix, I haven’t seen it successfully implemented aside from base OS monitoring.

    See my final post on Nagios for other alternatives, the reason Riemann wasn’t included is due to the complexity when layering it on top of other graphite based systems which are covered in the other post.

  • 141 Mirza   February 23, 2014 at 3:04 am

    While I have played with many monitoring distributions out there (Hyperic, Centreon, Nagios, Cacti, SCOM) I find Nagios to be the only tool that works and does not require a full time administrator (such as SCOM.. which is a beast).

    Have you heard of OMD/Check_MK and/or Ninja/Merlin by op5?

    We currently use OMD/Check_MK (which under the hood runs Nagios) and it does everything we need, the web interface is easy to use and you can get advanced network/server/service monitoring down tight in just a few days.

    The title of this article is very misleading.. sure if you are going with “plain” nagios it can be a PITA to setup but there are advanced implementations of Nagios using the Open Monitoring Distribution package and I think is second to none for getting the job done on monitoring/graphs.

  • 142 Jonah Kowall   February 23, 2014 at 11:03 pm

    Yes Mirza, I know OMD, Check_MK, Ninja and Merlin from op5. Nagios doesn’t handle basic topology or event suppression, it’s missing a lot of key features most monitoring tools have. The Nagios network monitoring capabilities are also very limited.

    Agreed we’re talking about open source Nagios, and not other solutions built on top of it. OMD is an improvement, but still not nearly easy enough.

  • 143 MaGGs   March 7, 2014 at 12:42 am

    Hi. Stupid theme. Maybe you never get in touch with real Nagios and / or Linux very deeply. Seems to me your technical knowledge is not a high level. I am working with Nagios for about 12 years now, less costs, no problems.

    Btw., Zabbix sucks. =)

    Greets, MaGGs

  • 144 Ken Merkel   March 17, 2014 at 2:56 pm

    I am amazed at the lack of research that went into this article. Especially since this is supposed to represent Gartner who is supposed to be an industry leader in IT.

    I work for a large organization and we have a very large installation of Nagios XI. When we compare the support we have received from the Nagios community to the support we used to receive from our commercial vendors…it’s uncomparable. The support is way better and the product does exactly what we want.

    We used to pay hundreds of thousands of dollars/year to monitor our environment. Now with the same numbers of staff we monitor our environment for $15,000/year and the end product is far superior to what any commercial product offers.

    Nagios is an industry leader in monitoring for a reason.

  • 145 Jonah Kowall   March 18, 2014 at 4:37 am

    Ken, you’ll notice this article is pretty clearly targeted at the Open Source products. There are links to my statements on the commercial products (such as Nagios XI). I advise clients regularly to take the path you have taken if they want the open nature of Nagios with support and additional functionality.

  • 146 Rich   March 19, 2014 at 2:05 pm

    The reality is that Open Source != Free. The $$’s saved on the bottom line are spent elsewhere as the intangible hours of the employee(s) configuring it. Nagios Core configuration is a skill, and it is a mistake by management to only have ONE person with that skill.

    Complaining that it doesn’t work and things don’t work as they should is simply from the author’s inexperience with the product and scripting. Nagios Core is very powerful, consistent, and, depending on the skill of the individual managing it, it can reach in to application to provide self-healing for predictable issues.

    Monitoring is an evolutionary process. Even out-of-the-box “inexpensive” tools require a fair bit of R&D, including learning the product, customizing checks, and reaching beyond the basic checks included with the product. It’s misleading to assume that a monitoring solution can be achieved by simply dropping in a product with nothing else invested. Doing that would set false expectations to the team and to management.

  • 147 Matt   March 21, 2014 at 1:47 pm

    OK, I read this and I read the comments and I have to agree with the majority. Nagios may not be shiny, but it does work and if use switch from Nagios to Icinga (especially Icinga-Web) then you can get the GUI-based point-and-click configuration that most of the “anti-” posters on this thread seem to be after.

    I’ve seen Icinga working across thousands of nodes with hundreds of thousands of service checks.

    I’m in the process of setting up a system that will scale to the same levels using check_mk and mk_livestatus, neither of which are difficult to setup.

    The Chef cookbooks that exist to auto-configure these things work incredibly well at scale – I provision a new node or instance, the next time chef runs on the icinga server it is automatically registered and monitoring starts.

    I have also configured Nagios to provide business logic and I would seriously argue that it is hard to write a new check or to do APM in Nagios/Icinga. As long as you can write a program that can be executed and exits with a meaningful message along with an error code of either 0 (OK), 1 (Warning), 2 (Critical) or (unknown) makes it incredibly powerful to monitor everything from webservices using cucmber-nagios through to call-waiting times in a call centre based on custom database queries.

    Nagios and Icinga are here to stay. It’s a shame people don’t realise this…

    (Disclaimer: Not a nagios employee, reseller or consultant, just a very happy user of the product and its forks and predecessors since it was NetSaint…)

  • 148 http://www.lyzka.krzywyraczek.pl   March 22, 2014 at 11:30 am

    Wow! This may am real solitary distinct of the most extremely caring blogs Weve actually make it all over with this area under discussion. On the whole Admirable. Im as well an expert in this area thus

  • 149 Jonathan Vella   March 27, 2014 at 11:00 pm

    To be fair it depends what you want to monitor as well.

    Have you come across https://xpect.io ?

    It’s a relatively new concept, which tackles monitoring events as opposed to states – such as payment, backups, scheduled jobs etc.

    Nagios is brilliant for state monitoring. What’s not so brilliant is the time effort it requires, especially if you are trying to monitor events. When you consider the cost of paying a DevOps to write scripts, it can easily become very expensive to maintain.

    Myself and a software architect wanted to build something which is more straight forward and allows DevOps and SysAdmins to focus on their projects rather than tackling monitoring ‘by hand’.

    The whole idea is that you set up an expectation, such as a ‘Database backup’ which should occur ‘Daily’. Xpect provides a code snippet (in multiple language options) which you simply paste in your code where the Database backup event completes. Done.

    As long as Xpect keeps receiving check-ins, everything is working as expected, if not you will be notified instantly. It can also integrate with Nagios, New Relic and PagerDuty for escalations.

    As a community of experts I would ask you to try it out for free, and see if it works for you. Any feedback would be greatly appreciated.

    We have also set up a blog post here: http://thoughts.disrapt.com/2014/03/13/architecture-of-xpect.html on how the architecture of Xpect is.

  • 150 Jonah Kowall   March 28, 2014 at 4:53 am

    Thanks for sharing, I haven’t heard of it before from clients, nor have they briefed me. The architecture you propose is pretty hacked together one could do much more easily with other tools out there if you want to put in the API work. I’d much rather use something like Data Dog for this type of use case, which provides more meaningful correlation, collaboration, and overall capabilities, but thanks for sharing.

  • 151 Nagios ServiceNow Integration - Evanios   April 11, 2014 at 10:29 am

    [...] free, and has a large variety of plug-ins available. Although it may lack some bells and whistles (Gartners words, not not mine), by pairing it with a solid Event Management solution it can be very effective. At the end of the [...]

  • 152 Kevin Rego   April 23, 2014 at 7:40 pm

    ” I haven’t heard of it before from clients, nor have they briefed me”

    Seems to be the only reply you have. Who cares what clients are saying, you’re supposed to be the expert.

    Perhaps if you were a better advisor you’d have done the research on how to implement Nagios/OMD/Icinga in tandem with Splunk or Logstash and Puppet or Chef.

    This would get them everything you’re proposing it needs, for free, and would give them the underpinnings of properly manageable and scale-able infrastructure.

    Poor article at best base on the rampant misinformation.

  • 153 Jonah Kowall   April 24, 2014 at 2:42 pm

    Kevin, on the open source side I have implemented all of these tools in my lab… When it comes to commercial offerings I want a briefing as there is more to buying a technology than the technology itself. Without understanding the business side of the vendor I’m not going to recommend relying on a vendor who may not be viable or have the right leadership.

    I’ve also covered the use of open source monitoring along with log indexing, but those are covered in other blog posts. Look at the in depth coverage around both areas within my blog and research.

    Thanks for your comments.

  • 154 Arun Nayak   May 12, 2014 at 9:09 am

    1.) Don’t write about something which you didn’t try.

    2.) Nagios is the Best tool and now I doubt about the infrastructure which you have used nagios for!!!

    3.) Come out of the research lab and visit Datacenters and places where nagios was used.

    4.) Nagios was not created to solve research problems, it was created for Monitoring.

    5.) Still if you are not convinced, better check your family doctor for mental issues.

    All the best.
    -AKN

  • 155 Jonah Kowall   May 12, 2014 at 4:15 pm

    Arun, I’ve implemented and managed Nagios installations in production environments, and I deploy it in my lab along with other tools and products I advise clients on regularly. I speak to hundreds of clients who use it and hear the same issues consistently. I appreciate your comments, but not your accusations.

  • 156 Gerry Johnson   May 14, 2014 at 10:41 am

    As I said in response to another of your rants, it just seems as though you are conducting a vendetta against Nagios.

    We use it here, monitoring several hundred devices and tens of thousands of services and it never goes wrong.

    Admittedly it takes some skills to set it up and develop new checks but I have worked in sites who have paid seven-figure sums for OpenView or Tivoli and need consultants on call.

    Then there are licensing costs which very quickly ramp up.

    Your aversion to Nagios does look very odd, and, judging by the number of posts not agreeing with you here, you are in the minority

  • 157 Jonah Kowall   May 15, 2014 at 2:38 pm

    Gerry I’m glad it’s working for you as it does for many people I speak with, but by and large over time the product has issues which are addressed by various add-ons to Nagios core. This is also the reason there are many redevelopment efforts which should fix a lot of these issues.

    I completely agree that “Openview” which is a really old name of the HP products or IBM Tivoli have even larger issues, and come with bigger problems for organizations. Those suites of products are far more deep and wide than Nagios which is a server monitoring technology. The costs not only in terms of implementation of these large legacy toolsets are problematic (although both companies are working hard on easy to implement and lower cost solutions) but the ongoing maintenance is a problem as well.

    I have no aversion to Nagios, the community is excellent, but with a 14 year old technology it’s been passed, similar to your mentions of legacy tools from HP and IBM!

    There is no need to agree, this is a similar issue to the everlasting Mac vs Windows vs Linux argument, everyone is passionate, and I encourage and participate in dialog.

  • 158 Zoe   May 26, 2014 at 7:02 pm

    I lead a team of Nagios developers where we administrate more than 30 Nagios implementations, some of my customers have more than 50K nodes, we had overcome the issue with checks just using vmware and putting in each virtual machine the correct amount of nodes supported for a 5 minute checks, and concentrate everything in a dashboard using MTOS. I can say that anybody that understands a little bit of linux can handle Nagios (BTW Nagios core is way better than XI).
    Cheers,
    Just another linux lover…

  • 159 Jonah Kowall   May 27, 2014 at 2:26 pm

    Zoe, thanks for your comments. 5 minute polling is likely to miss a lot of issues unless they are major outages.

  • 160 OSSfail   July 1, 2014 at 7:28 am

    This isn’t even an article about Nagios, it’s an article about open source in general. OSS is great when there is an expert resource available, in all other cases the countless payroll hours to make things work far exceed any cost savings of not paying for a commercial solution with commercial support.

  • 161 Guy Boisvert   July 4, 2014 at 5:45 pm

    Re: OSSFail

    “[...] in all other cases the countless payroll hours to make things works far exceed any cost saving of not paying for a commercial solution with commercial support.”

    This is a broad statuement to say the least… Would you share with us the data and facts that your opinion is based upon? This is much too easy IMHO to pitch broad statements like this without giving any proof, study, experiences, etc.

    Business cases can vary a lot even inside the same company considering a lot of metrics… Commercial Softwares need experts too and speek to any Windows SysAdmin, it’s not just about clicking in a familiar Windows interface… Even basic Windows domain administration takes knowledge. Knowledge is mandatory! Inhouse or as an external ressource…

    As for Nagios, it’s very flexible and will do great if configured / used the proper way . And ask the expert that configure it to document, it should always be. I’m a 25 years IT veteran and i can assure i saw nighmares, both OSS and Commercial! It’s almost always a matter of knowledge and selecting the right tools for the job. There are many metrics (not in order):

    – What you want to achieve
    – What ressources you have at hands
    – Costs
    – Evolutivity
    – Adaptability
    – Stability
    – Integration in your environment
    – Security
    – Etc.

  • 162 Jonah Kowall   July 6, 2014 at 10:08 pm

    Good points here Guy, but I hear time and time again from clients who’ve fallen victim to the issues of the hidden costs of free software. Many can make it work, but at scale and increased complexity a dated monitoring tool like Nagios often causes issues.

  • 163 Oliver   July 17, 2014 at 7:50 am

    Nagios is a great Framework to make monitoring things easier and less complex if you spend some time which you always should!
    Even if you buy a commercial solution and give your experience/knowledge to another company.

    We’ve some business processes which are visualized with Nagvis and some custom written plugins if we had to integrate these bps into closed monitoring systems we would have spent a lot of money..

    But i agree if you dont have man power to install, configure and maintain nagios, you might use some propietary product or you buy some support for integrating nagios as there are plenty of service providers for nagios/icinga which also support you in integrating your bps..

    I would agree Guy: it highly depends on your needs and desire

    Greetings
    Oliver

  • 164 Gartner: Got Nagios? Get Rid Of It!   August 11, 2014 at 5:26 am

    [...] Wrote Jonah Kowall of Gartner, [...]

  • 165 Nick   August 28, 2014 at 12:03 am

    I seems to me that this article is not very well documented. I’m saying documented since it’s clear that Mr Jonah Kowall doesn’t know what he is talking about.
    Not that Nagios is the only way to go, but stating stupidities like “Nagios with agent architecture (SNMP can scale better) does not scale” or “the check architecture just cannot handle volumes” and “I’m sure the large implementations have done the following:

    1. not using agents with checks
    2. changed the code to optimize it
    3. are using a lot of hardware to scale the solution”.

    Seriously, it’s gold, it’s like getting it wrong all the way ! So this article IS a troll, and I can understand the upset answers of people probably knowing what they talk about.

    I have a big experience on Nagios, it does have his problems, but none are mentioned here.

    Also please, I would be happy to know the products names you are advising your customer to use for $2-$20.

    Last word about the auto-discovery, it’s a nice option but more for messy infrastructure that are not able to have a proper count of they hosts (CMDB anyone ?). So either messy places, either tiny place, but certainly not big, otherwise they’d be dead by now …. (and I’m talking with experience here !)

  • 166 Jonah Kowall   September 2, 2014 at 3:02 pm

    Thanks for the reply Nick.

    Nagios agents are completely inefficient, this is part of the reason for all the monitoring rewrites happening right now. After speaking to hundreds of monitoring users in the last year, managing agents in general… Inclusive of patching, configuration, troubleshooting, and configuration is time consuming. Moving towards agentless monitoring is clearly the right direction for almost all users. This is not always possible, but is normally the right way to handle monitoring of system level metrics. I speak to many large Nagios users who have issues managing it at scale. Essentially it’s used for basic system metrics, and most progressive organizations have moved towards using streaming collection engines instead of Nagios (collectd)

    I’m completely against the notion of a CMDB the way they exist today, but that doesn’t mean that discovery in a more lightweight manner isn’t the right way to insure proper monitoring coverage. If you do not do discovery in most organizations you will always have coverage gaps.

  • 167 Arie   September 3, 2014 at 2:46 pm

    As mentioned before, we are monitoring with a Nagios/Check_MK combination, and we can see/monitor a lot of our needs but there is a lacking me when it comes to application monitoring in a different way.

    A movement that is going around as I sense it is the upcoming use of log management tools like ELK/Logstash/Graylog2 to fill up this gap.

    These tools are making it possible to me to fill up the gap I was having, and log management is complementary to monitoring with Nagios/Check_MK.

  • 168 Jonah Kowall   September 3, 2014 at 8:32 pm

    Completely agree Arie, ELK has helped deal with some of the issue of Nagios (See my blogs on ELK). Managing large scale Check_MK is still archaic at best. Without policy management or a proper configuration/discovery system it’s hard to scale.

  • 169 hotmail customer service number uk   September 23, 2014 at 4:28 am

    The assistance from the hotmail customer care become important for
    users System.Drawing.Bitmap out the users who are facing difficulty with using hotmail
    and searching for right kind associated with support.

Leave a Comment