by Jonah Kowall | November 19, 2014 | 2 Comments
We’ve published new research on mobile application performance monitoring (APM), this research highlights some of the changes in the market over the last year since we published our first notes a year ago. There have been many interesting developments, along with a growing interest among our client base. Some of the changes include the melding of analytics technologies with APM visibility to create stronger platforms to understand end user experience and deeper device and application monitoring. We also analyze behaviors which cause user abandonment, and why quality is an issue within mobile applications.
Additional analysis of the buyers, and future state of the market are included in this research. We are planning on releasing a market guide for Mobile APM in Q1, which will be an update to the vendor landscape we did in late 2013. That research will also include market sizing estimates including usage, penetration, and revenues.
I will be presenting a brand new mobile APM content at the upcoming Gartner Data Center conference which complements this research note (http://www.gartner.com/technology/summits/na/data-center/).
Category: Analytics APM IT Operations Mobile Monitoring SaaS Tags:
by Jonah Kowall | November 15, 2014 | 4 Comments
Sorry for the delay, it’s been crazy for the last week to say the least. Last week I was able to attend the inaugural AppDynamics user conference, AppSphere. It was a short trip, I was unable to attend for the final day which included interesting breakout sessions (seems to be a theme the second half of this year).
AppDynamics made several key announcements during the conference including great revenue and employee growth, it’s no question this high flying company is on a path towards IPO as another competitor is currently doing. The main announcements included the December availability of AppDynamics fall 2014 release, which includes a rearchitecture of the platform, dubbed “Application Intelligence Platform”. They also announced the availability of over 100 platform extensions to extend and enhance the value of the platform. The kickoff keynote by founders CEO Jyoti Bansal and CTO Bhaskar Sunkara also announced the new ‘Application Analytics’ module (pricing has yet to be determined), a core capability to extract, correlate and and visualize ITOA and business impact data, also available in December. This underpinning architectural change in the product will allow them to do much more sophisticated and diverse data collection and larger scale analytics (web-scale). (Related research published not long ago Monitoring Must Evolve to Meet Tomorrow’s Demands -http://www.gartner.com/document/2809724).
The other big expansion was the identification of the need to address gaps such as a broader unified monitoring platform. This is slightly different than Gartner’s current perspective on Unified monitoring, which is availability focused, AppDynamics has much grander plans for the platform (Client research : Modernize Your Monitoring Strategy by Combining Unified Monitoring and Log Analytics Tools – http://www.gartner.com/document/2615618).
Additional demos and announcements included a virtual war room feature which tries to replace the painful way we troubleshoot today, often with long and painful conference calls and meetings. This has been attempted before, unsuccessfully, but I wish them luck One announcement, lacking detail, was a beta version of the AppDynamics synthetic monitoring product. Synthetic monitoring is complementary to APM, and we’ve seen more of these products released within existing APM offerings in the past few months including (ex: AlertSite and New Relic); it will help remove market fragmentation and provide new use cases for synthetic monitoring.
Towards the end of day one I was asked to moderate the cloud panel, which was an interesting and lively discussion. While I would have liked a bit more bloodshed, it was interesting and I learned quite a bit myself from the participants. There was an interesting mix of vendors on the panel from IBM (SoftLayer), Redhat, Microsoft, Pivotal, and Google. Each had different perspectives on cloud (public and private), the way people were using it, and where it was heading. More detail of this conversation can be found here : http://www.appdynamics.com/blog/news/appsphere-speaker-panel-cloudy-with-a-chance-of-innovation/
I had several other meetings during my short stay, which were all engaging and interesting. There were many good customer sessions within my formal and informal discussions, these passionate customers were solving very real and complex problems using AppDynamics (in addition to other vendor) products. It’s great to see customers using APM technologies to solve business problems. Kudos to the AppDynamics team for a great conference.
Disclosure : AppDynamics paid Gartner to have me keynote and participate in the conference.
Category: Analytics APM Big Data DevOps IT Operations ITOA Logfile Mobile Monitoring SaaS Trade Show Tags:
by Jonah Kowall | November 6, 2014 | 8 Comments
Unfortunately we’ve lost one of the innovators in APM this year already (Optier), and we have another trying to survive. When a company creates and builds unique and innovative technology that doesn’t equate to having a well run a successful business. Business execution encompases many facets outside of technology to build a successful company. About 2 weeks ago I was made aware of some issues at APM vendor Correlsense, before lashing out that the sky is falling we did some due diligence. You can find a couple of stories here in the press (which are not entirely accurate):
http://www.calcalist.co.il/internet/articles/0,7340,L-3643229,00.html (this one is hebrew, but google translate handles it fine)
Taking a step back, Correlsense has unique technology and has been around for a while. They have raised funding through several smaller rounds (http://www.crunchbase.com/organization/correlsense). In the latest change of management about 20 months ago they had brought in a veteran management team who came from very traditional enterprise software businesses. Most of the readers here realize what we’ve been writing for a while now, APM has undergone a radical change over the past few years, those which evolve win, and those which do not lose over time. The enterprise APM market is not the same as the enterprise monitoring market of 5 years ago. Things have changed, and the management team at Correlsense came from classical backgrounds, and refused to evolve (basically ignoring people like me, and believing they were correct). Correlsense was once in the APM Magic Quadrant, but they could not deliver a SaaS based product, hence when we made that a requirement in 2013 they were unable to qualify. The management team determined there was no demand for SaaS by the enterprise, and they were not going to put that on the roadmap. This was a grave mistake (one which Optier realized too late given the funding history) and prevented them from adapting to market changes. Additionally about 3 years ago with the massive growth in environments Correlsense had some scalability issues. These were solved with a large rewrite of the product on a modern non-relational data store. That is very hard to accomplish, and the engineering team deserves kudos for pulling that off, it was not without some bruises. Regardless of the unique and differentiated technology, the business execution was a major issue which did not change.
The result of this is that Correlsense has had layoffs, especially in the US headquarters, they have also moved away from the sales and business models which were created by the previous management team. The company has moved it’s headquarters back to Herzelia, Israel (where it came from) and they are still sorting out the CEO. Hopefully it will be one of the founders as he knows the market and the technology of APM. The result is a change of focus, away from the US and towards Europe where they have seen momentum. Customers in the US will still have support, and there will be a smaller presence in the US. While I don’t agree that Europe is a particularly strong market with the fragmentation, a good sales and support team can grow a business. They will be focusing on partners and channels to handle most of the sales versus a centralized sales model. They have a common investor with Centerity, and that partnership makes a lot of sense to grow both businesses. Centerity is a unified monitoring product we have covered in other research.
The company plans to build a SaaS platform, with a try and buy model. This will be the focus of the future company direction. The timing, naming and functionality have yet to be determined. They will focus on usability, and a modern application consumable by tablet and mobile devices. They will move away from large enterprise software deals towards lower cost subscription pricing. This is a major shift which many have tried, but we’ll see if Correlsense can pull it off. They will still have the current SharePath Enterprise product for focusing on difficult to monitor applications such as thick clients, Tibco, Tuxedo, WebSphere Message Broker, Oracle Forms, and other legacy applications. This is clearly not going to be the focus moving forward, but is of use to some buyers. With the combined IPM and APM capabilities they certainly have great intellectual property and a possibility to correct the course of the company, we wish them luck!
Comments below, or on twitter @jkowall please.
Lanir Shacham who is the founder is now the CEO as of two weeks ago
Category: Analytics APM Big Data IT Operations Monitoring SaaS Tags:
by Jonah Kowall | November 3, 2014 | 2 Comments
Building large research projects is a major undertaking, not just for the analysts, but the vendors who provide large amounts of data we request, and references we leverage aside from the well over 1,400 APM inquiries we take yearly from end user clients buying and implementing APM. We’ve published the 2014 edition of our yearly magic quadrant for APM. In fact it published a week ago, but I’ve been too busy to blog about it. Understand the technology cut off was in the summer, hence enhancements since then will be counted in the 2015 research.
There are already several press releases and licensed copies available from some of those press releases. We have three leaders in this research, and they have been extending their ability to cater towards APM specific buyers. These three leaders AppDynamics, Compuware (soon to be Dynatrace), and New Relic are moving in different directions, but using their depth in application instrumentation to lead the efforts.
We had one new entrant, SmartBear, who acquired Spanish APM company Lucierna and created a new interesting offering called AlerSite UXM. I have been testing the instrumentation in the lab recently, I will be bloging on it shortly. There have also been major shifts in the last year including CA, HP, IBM, and Microsoft. You can read about these changes, trends, and insight in the research.
There will be a new research note called a Critical Capabilities coming out in the next few weeks which will be a yearly companion to this Magic Quadrant. We are doing more of these across Gartner as they are very complementary in nature. The Magic Quadrant rates and evaluates a vendor’s ability to execute and the vision. The Critical Capabilities document (in our case) will outline several typical buyers or communities of APM users, and show the vendor’s ability to meet specific use cases. The use cases span typical and emerging criteria based on technologies and capabilities, these each have different importance to each buyer. Although I’d like to give away more than I have about this exciting upcoming note, I will hold off for a few weeks :). Think of this document as a technical and deep view of the vendor’s ability to meet these capabilities.
Please leave comments, thanks for reading, and we appreciate the vendors who participated in the research process.
Category: Analytics APM Big Data IT Operations ITOA Mobile Tags:
by Jonah Kowall | October 24, 2014 | 2 Comments
I was invited and able to attend Compuware Perform in Orlando the week of November 6th for a couple days before heading off to Europe for 10 days of vacation, hence this blog post was slightly delayed (not to mention the Magic Quadrant and other research which is imminent now). The conference was great to attend as Dynatrace re-introduced the world to the brand, what old is once again new. The products have been renamed as follows:
John Van Siclen who was previously the CEO of dynaTrace (acquired by Compuware in 2011) was the general manager for the Compuware APM business. He is now the General Manager for the Dynatrace company. The expectations are that this will be a standalone company when the privatization of Compuware closes in the fourth quarter of this year (more on that later). John’s keynote had some interesting points, Compuware has an impressive 89.9 net promoter score, and an online community of 84,000 people.
Dynatrace launched 4 key targets and messages for the solutions : Launch readiness, User analytics, Performance Engineering, and Production Monitoring
Dynatrace plugins will be open (as other APM companies have done) and they will be leveraging github : http://www.compuware.com/en_us/application-performance-management/products/user-experience-management/real-user-monitoring-web-and-mobile/plugin.html
Innovation will flow from Ruxit to Dynatrace including the network probe agents and the UI. I’ve already posted my thoughts and time with Ruxit, which were positive.
Dynatrace 6 was unveiled which included agent support for technologies such as NGnix, IBM IMS. There were improvements for TIBCO ActiveMatrix, TIBCO EMS, Java 8, iOS 8, NGINX, HBase, Cassandra, MongoDB and iPlanet. Dynatrace showed the new Web UI, which provides some high level views outside of the thick java based client they use today. I wasn’t too impressed with the new dashboard and UI, but it was an early-state prototype I wouldn’t have expected to see in a public forum. I am sure that will improve considerably before it ships next year.
Some nice improvements in the synthetic monitoring, including a free web test for lead generation (I still believe what you get for free with webpagetest is far more complete). Investment in synthetic monitoring seems too heavy, this has been an issue with Compuware for quite some time.
I spent time in a couple sessions on DCRUM, and the product are still very network buyer and legacy application focused. I wasn’t too impressed with what I saw. I would have liked to have seen a more innovative approach to solving these problems, similar to the agent present in Ruxit, which is how we see the future of network analysis being done.
The one large piece missing from my discussions and the messaging was analytics, while Dynatrace has lots of great analytics technologies within the products for determining root cause, understanding performance deviations, and finding the major issues within applications, the broader analytics and ITOA strategies many are strategically invested in is something which Dynatrace’s messaging consists of integrations with providers like Geckoboard, Splunk, and others. This is clearly not what APM buyers are looking for today.
On the non-product side many of you are aware that private equity (PE) firm Thoma Bravo, who have an extensive background in the monitoring and management space (Network Instruments, Keynote, Infovista, and others) have decided to purchase Compuware and divide the company into several parts to drive growth. While many PE firms operate in other ways Thoma Bravo is a unique firm. One of the VPs Chip Virnig spoke a little about how APM is a great market to place bets (which I agree with) – http://www.thomabravo.com/team/virnig/
Some recommended reading about PE firms (sorry clients only) : How to Re-evaluate Strategic Vendors Acquired by Private Equity
I spoke to several customers, and some were using end to end mobile APM. Interesting use cases, and capabilities within the offering which have matured nicely. Expect new research and presentations on mobile APM in the next 6 weeks at our upcoming Gartner Data Center Conference. Dynatrace 6 offers major improvements on scalability in terms of how many controllers are needed. Customers confirmed they needed far fewer controllers and management servers in the newer products, which is good to hear from end users. There is more planned in terms of scale and capabilities.
Only being able to attend one day of three, I got a lot of value out of this conference.
Disclosure : Compuware paid for travel and hotel for my attendance at this conference
Category: Analytics APM IT Operations ITOA Mobile Monitoring Tags:
by Jonah Kowall | October 23, 2014 | 4 Comments
I’ve written quite a bit about how analysts work, and here are some of my personal favorite reads, but they are more focused on vendors versus end users or investors:
I’ve shared my personal social media policy, along with other personal ways I do work.
This blog post is specific to attendance at vendor conferences, for transparency. Some of the conferences I attend I am provided media access and use my credentials as someone who covers and writes about a specific vendor or market. When there are vendor conferences they often pay for travel and hotel for the conference should I choose to attend. Similarly they can also hire Gartner analysts to speak at conferences, which requires a larger commitment of financial investment on their part.
In addition to attending the conference I also meet with other clients, prospects, and interesting people at conferences, so sometimes Gartner will pay for these conferences for research, or to support our business growth and success.
In my past blog posts, I have not been completely transparent around my conference attendance, and it hasn’t been due to wishing to avoid being honest, I just hadn’t really thought about it. When reading some of the other coverage in the press, I soon realized that in order to be fair, I needed to do the same thing.
Going forward you will see full disclosure of when conferences were paid for by a vendor, or when conference admissions were provided by another organization.
Category: Trade Show Tags:
by Jonah Kowall | October 7, 2014 | 4 Comments
Compuware APM (now Dynatrace) has a new product and brand they are treating internally like a startup. This new software product has been in the works for a while and takes a unique approach to APM. Although they would like this to be separate from Compuware, it’s not quite the case.
They are clearly targeting those web first APM companies (such as New Relic and AppNeta). Additionally with these unique features and ease of use focus it will appeal to enterprises, the same way New Relic has been making inroads around enterprise adoption. Come and hear more at the upcoming Gartner Data Center Conference in December where we’ll have end user speakers talking about this!
The product is SaaS only and they provide a free trial. The billing is usage based (hours), there is not a “monthly” price aside from paying for the hours in a month per instance being monitored. Upon signing up the setup of the server side takes a little time to provision. Once provisioned you login (below):
The main login allows provisioning and management of multiple environments (essentially AWS zones today, but in the future that could be a lot more). The problem is there isn’t really a wizard, you just get dropped in (the Ruxit team has provided feedback that they have already improved this since my testing, don’t you love the iterations with SaaS):
After clicking the link for the proper environment, a wizard is presented helping the implementation of the host agents:
Download the agent, which is already keyed for your environment, this makes the install very easy. On the Linux side it can be downloaded, or use the wget link is provided. In this example agents were installed on Windows and Linux, this is the Windows download link.
Install completed on the server, and you have my tenant ID, but whatever :):
Ruxit is very easy to use and implement here is a description of what is supported on Linux with the product:
Once installed (for example on my Linux host) it discovered and started monitoring the Tomcat install which is running on the JIRA host. What makes Ruxit unique is there was no editing of scripts, configuration files, or anything. The product really implements completely seamlessly. This is a new trend you will see among leading APM tools, but Ruxit is first to market with this feature:
On the windows side it discovers multiple services. Ruxit doesn’t go as deep on .NET as it does on Java right now:
More screenshots on a SQL server:
Once monitored here is the dashboard. The tiled interface is scrollable and all HTML5. It feels and acts like Windows 8 in terms of the UI:
The dashboard is customizable, sections can easily be added or organized via drag and drop:
When adding a tile here are some of the options:
One of the unique elements of Ruxit is that it has some IPM capabilities, hence it goes deeper into the infrastructure and topology of applications outside of what is instrumented. The Smartscape view is one representation of this. Layers represent the applications, services, hosts, and data centers (support for VMware APIs is supported) on the left hand side. You will see additional support for IaaS and virtualization providers in the future:
Clicking an entity provides details of the layers within the application (this one is Sharepoint – .NET).
Here is the view for a Java App:
When drilling into the component high level views are presented including metrics, methods, problems, and events. Rich data is presented, especially in Java applications. Less detail is shown for .NET or other applications (The Ruxit team has clarified PHP and .NET will be enhanced this year, and node.js is currently in beta):
Yet another view shows the services across tiers, you’ll notice the breakdown on the left hand column:
Here is the same view for the Sharepoint app on IIS:
Similarly here is a database view to give you the depth via JDBC:
Drill into the database activity, this shows the breakdown of slow statements and where time is being spent:
Backing up to the Smartscape, there are interesting views from the database host perspective. Here you can see that both the .NET and Java (Sharepoint and JIRA) are both using the SQL backend.
Sticking to the host level view, system level metrics are visible when drilling down:
Clicking on a metric shows graphs and other data associated, in this case memory:
On the SQL server this is the breakdown of processor usage by process:
Here is the Disk IO graph:
The ever important disk latency broken down by reads and writes.
Process level view of the SQL processes and associated callers:
Here are some alerts, and a new take on alerting:
Some RUM screenshots:
Ruxit is a unique and much needed entry into the APM market, with an easy to implement SaaS deployment model providing breadth and just enough depth for many buyers. The product has unique and interesting IPM capabilities which span outside of the traditional APM market, which tend to focus on visibility and instrumentation within the applications themselves. We will see more similar products begin to emerge along with combinational products of unified monitoring and IPM to bridge availability and performance.
The pricing model is different than many other products with $.15 per hour for application monitoring, and $.15 per 1,000 user visits for RUM. This works out to about $108 per host per month, and the RUM is priced based on volume. This is competitive pricing, but more depth is needed, especially around RUM (The Ruxit team says this is an area of focus for this year).
Will be interesting to see how adoption goes of this new and fresh APM product.
Category: APM IT Operations Monitoring Pick of The Week SaaS Tags:
by Jonah Kowall | September 17, 2014 | 4 Comments
In order to assist our vendor clients we often do strategy days, these are typically full or half days where I work with clients to help them craft strategy, go to market, or other consultative help we cannot accomplish during our 30 to 60 minute inquiry phone calls. Our vendor clients find these very helpful, hence we have a good amount of demand for these sessions. The problem is when analyst resources are tied up for extended amount of times it causes longer wait times for phone calls due to travel and other requirements.
Creating Magic Quadrants is a very lengthy process, which consumes a lot of analyst and vendor resource and time. We regularly include vendors who are not clients in our Magic Quadrants and research, and we provide the same level of analyst access to non-clients as clients during Magic Quadrant processes. Additionally we do not discuss Magic Quadrants with clients or non-clients once they are in process, aside from during specific conversations which are focused on the Magic Quadrant.
The way we handle vendor interactions during Magic Quadrant development is just one example of Gartner maintaining independence. Gartner analysts also don’t deliver white papers, or other vendor sponsored research notes, and analysts are also not compensated or otherwise tied to the sale of our products or consulting. These are reasons why Gartner is differentiated and how we avoid our bias.
Avoiding bias is critical for us to deliver the best product possible, this is always the case. This can prevent analysts from doing SAS with vendors in a specific Magic Quadrant during authoring.
I always strive to be transparent, and thought this would be good to share with the public and those I enjoy helping build the best technology and make the right decisions.
9/22/14: A few edits as requested by David Black, who’s a VP in the Content and Methodologies area. Expect more news around this in the coming weeks from Gartner across the company.
Category: Uncategorized Tags:
by Jonah Kowall | September 16, 2014 | 2 Comments
Former 2013 APM cool vendor Renesys was recently bought by Dyn http://dyn.com/blog/dyn-acquires-renesys-the-global-authority-on-internet-intelligence-2/. You would not be too far off in saying Dyn, who’s known for DNS services, traffic management (using DNS), and email marketing software (via acquisition) is making a move into a new market. The interesting thing as we highlighted in the cool vendor report is Renesys has a unique data set, you can say the same thing about Dyn who manages traffic, email, and DNS for a lot of companies out there here are some interesting stats (sorry I like data too much)
- 327k queries per second (monthly average)
- Over 1 billion emails sent per month
- 230+ countries & territories occupied by Dyn users
- +65k Domains Registered (2013)
After buying Renesys they began to integrate and analyze some of this vast amount of data into a more operationally focused product. Here is Internet Intelligence, which was released today : http://www.digitaljournal.com/pr/2190339. I was kind of upset they put me on a trial since I wanted to have this tool after using it for a bit. After the login here are the highlight of the product:
As you can see the focus on the initial product are focused on diagnostics and measurement, and less so on monitoring, but some elements come out later. This next screenshot shows the visual connectivity between the AS numbers.
As you can see the focus on the initial product are focused on diagnostics and measurement, and less so on monitoring, but some elements come out later. This next screenshot shows the visual connectivity between the AS numbers.
Additionally you can see them visually on a map and explore them in an easy manner:
Here is a view for this specific AS number:
You can also see the vantage from a point and the latency as you move from that point, this allows you to explore performance in a more ad-hoc manner:
A path can be selected and explored as well.
Taking these two views together you can look at path and latency data:
Latency can be examined over time as well, which can provide an indication of what the network latency might look like to a customer on a specific ISP. You can see how this would be useful for those hosting services or running SaaS businesses:
If looking to improve performance to specific customers or areas you can also compare and understand providers and options connecting points:
I found the measurements showing variability to be particularly interesting, the use of the scatter plot shows you the actual measurement variance:
Looking out over longer periods of time you can see more about path variability:
This is about the only monitoring in the product today, I expect this to improve a lot. You will notice the network events bulletin, they have an rss feed for this data, which I actually have in my feedly and love!
In order to setup monitoring you must add the assets into your portfolio:
As you can see or at least imagine, this is an interesting but early product. It’s good to see this great data being exposed for operational purposes, and I expect it to become more relevant as the internet is so critical for almost every business today.
Please leave comments and questions here, or via twitter @jkowall
Category: Analytics APM Monitoring Pick of The Week SaaS Tags:
by Jonah Kowall | September 9, 2014 | 3 Comments
I’ve been eagerly awaiting the publication of this new research. I worked with colleagues from other coverage areas including Magnus Revang, who focuses on application development (UI/UX), Magnus did most of the work on this note and really drove it. I helped with contributions and work, along with Ray Valdes who covers web technologies and social platforms. The result of this was a really interesting research note we pull in a lot of research around performance and speed, and how this is related to your bottom line. Additionally we start to dig into why performance is critical when it comes to user experience, and the relationship between efficiency, engagement, and performance. We then move towards the recommendations as to how to measure, improve, and iterate on performance improvements including a lot of free and commercial tools to help with these steps.
Clients can find the research here :
You can follow and interact with the authors on twitter:
Magnus Revang @MagnusRevang
Ray Valdez @rayval
Jonah Kowall @jkowall
Category: Analytics APM Monitoring Tags: