One of the highlights of 2017 for AppDynamics was AppD Summit New York. We took over NYC’s Pier 36 with over 1200 attendees and made big announcements around the next generation of Business iQ, new machine learning capabilities, and upcoming IoT and network visibility.
We were also able to share the spotlight with some amazing partners, including Evolven Software, who was an exhibitor at the event. We’re excited to share this Q&A with Evolven’s CEO, Sasha Gilenson, as we reflect back on the event.
You say that “change is the root of all evil” – what do you mean by that?
SG: Gartner researches state that 85% of performance and availability issues are caused by some kind of change. These can be changes in configuration, data, code, workload, and infrastructure. But there is often a change that triggers an incident, unless it’s a hardware failure. Evolven customers also confirm these statistics. Everyone agrees that one of the first questions in an incident war room, is “What changed?” The challenge is that many of the changes are unknown to the IT organization. Identifying them and then deciding which change is a culprit takes a lot of time and effort
What value do you deliver to AppDynamics users?
SG: AppDynamics provide a fantastic solution to detect performance and availability issues and to narrow down location of the issue to the exact service call, function or query taking too much time or failing, or in other words – AppDynamics tells you “what’s wrong”. However, in many cases the “what’s changed?” question still stands– why is this function or query is slow now when it worked fine just a few hours ago. Is it a change in the function itself, in the call path, in the data the function expects, workload, or environment configuration? Evolven automatically detects all these changes as they happen and correlates them with AppDynamics findings (health rule violations, performance KPIs, transaction topology), and identifies changes which are the most probable root cause of the incident.
We asked Jonah Kowall, AppDynamics’ VP Market Development and Insights, to share his insights on the joint value proposition.
Jonah: We’ve seen different practices within the silos of IT Operations around availability and performance as a practice and change and configuration management as a practice. Often times configuration management is tied to either security or other initiatives. The result is that monitoring tools can isolate a root cause of failure or slowdown from a technical perspective, for example poorly performing code, slow sql, malformed data coming into an API call, failed hardware, or degraded hardware. By working to combine enterprise actual changes as well as configuration management and APM that isolation can be done to the level of what changed, who changed it, and what was the linked change management ticket, build, or release. This is much more powerful, especially when the technology can support both legacy or typical enterprise technologies and modern technology stacks.
What is unique about Evolven and its approach?
SG: The main goal of IT is to deliver and operate rapidly, safely and securely changes supporting organization’s business requirements. To support this goal IT implements tools to manage and automate IT processes which drive change, along with tools to monitor symptoms of issues resulting from wrong by changes. However a critical component is still missing today – a technology for detecting and analyzing what actually changed. Evolven Change Analytics is the only solution that automatically detects actual, most granular changes across end-to-end IT environments, estimates risk of these changes and correlates them with investigated issues. And the Evolven’s solution does all this in data centers, virtualized environments, private and public cloud
The Summit featured a session on machine learning (ML). How does Evolven use ML to benefit its customers?
SG: Machine Learning is the heart of Evolven’s analytics engine. Thousands and thousands of granular changes are implemented on daily basis in IT environments. Just knowing what changed is not enough to prevent incidents and problems and investigate them faster. IT specialists cannot spent their time going through all this data looking for relevant information.
This is the goal of our machine learning analytics – to identify the changes that risk performance and availability of business systems and directly point to the changes that caused issues that need to be investigated. Evolven also uses machine learning for correlating the changes we detect with other data from existing IT tools, like AppDynamics.
Can you share some examples of how existing customers are using Evolven?
SG: There is a wide range of user scenarios our customers follow applying Evolven in their environment. However most of them use Evolven to accelerate issue investigation and to prevent issues. For example, we have numerous customers using Evolven to analyze consistency of the application and environment configuration across load balanced and clustered servers. Detecting and remediating risky inconsistencies improves system reliability and resiliency. Let me illustrate incident investigation process using the integration between AppDynamics and Evolven with a few technical examples of real issues solved by our customers.
Example 1:
AppDynamics detects a transaction responding in less than 5 seconds. It rapidly narrows the issue down to a JDBC query that looks legitimate. Without Evolven you will need a DBA to debug the query. The DBA might need to evaluate a schema, database configuration and even environment configuration. This can take a lot of time.
But the query has worked before. What changed?
Evolven instantly presents a change deployed last night in a stored procedure that is invoked by the query. A developer for some reason added an option to recompile one of the statements in the procedure. The issue was not detected in testing as test data sets are smaller than the production ones.
Example 2:
AppDynamics detects that a significant portion of users executing a specific transaction wait for more than 10 seconds. AppDynamics points to a third party component that slows down the transaction. The component is a black box. Without Evolven production support needs to escalate the issue to a vendor supplying this component.
But only part of the users experience poor response time. What is different?
Evolven instantly compares virtual hosts running the third party component and detect a difference in the patch level of .Net framework across the servers. The component relies on .Net. Obviously some of the virtual hosts are still based on an older image that was not patched.
What was your experience from the NYC Summit?
SG: It was a very unique event in many senses. The open space layout of the venue enabled productive and quite intensive interaction between the attendees, the AppDynamics team and partners. like Evolven. The list of attendees and companies they represented was impressive. This was the first time Evolven was able to introduce its unique approach, technology and value of the integration with AppDynamics to such a wide range of AppDynamics users. I was excited to find that anyone that spoke with me or my team confirmed the value and relevance of our solution and the integration for their organizations.
Check out highlights of AppD Summit New York here.
