CA APM Review

Allows us to pull metric data from other monitoring tools. Dashboard tools need a significant update.

What is most valuable?

  • Deep-dive application instrumentation allows us to diagnose application performance and availability problems.
  • Powerful querying capabilities of the metrics data.
  • Ability to build custom dashboards and reports.
  • Ability to create custom metrics using calculators or custom JavaScript calculators.
  • Ability to capture and feed custom metrics into CA APM. This same capability allows us pull metric data from other monitoring tools and feed that into CA APM to provide that single-pane-of-glass view.
  • Metric Data API that allows querying and extraction of metric data.
  • Java and .NET application monitoring.

How has it helped my organization?

It has provided application teams with another tool in the tool box to diagnose complex performance and availability issues. It also allows for the early detection of an issue before we hear about it from our customers. CA APM reduces our time to resolution and improves our customer confidence.

What needs improvement?

While calculators are present, they don’t support the ability to create a custom calculation on a metric. To do this, a custom JavaScript calculator must be built.

The dashboard functionality is what one might expect from the early 2000s. The dashboard tools need a significant update in terms of the widgets that are available, usability, and ability to modify how data points are labeled.

The custom report functionality also needs to be updated with the ability to apply different widgets, rather than just a line graph, and the ability to modify how data points are labeled.

The capability to execute a shell command on an alert condition is present. However, the implementation makes it difficult to find potential use cases. In our environment, it’s unfortunately something we are unable to make use of, at least directly. In short, CA APM is not a tool I would recommend for implementation of self-healing. However, it can be used to automatically advertise that a condition under which self-healing should be kicked off exists.

.NET agents are not at par with Java agents and there is a gap that needs to be closed.

For how long have I used the solution?

We’ve been actively using CA APM 10.2 in non-production and production environments since March 2016.

What do I think about the stability of the solution?

We have had stability issues, but we have worked out the major issues with CA and are working towards resolution of the minor stability issues. The fixes for the stability issues we experienced have been implemented in versions of CA APM following APM 10.2.

What do I think about the scalability of the solution?

We have not had any scalability issues. I will note that it is important to have an understanding of how APM works and, if you’re new to the product, to work closely with CA on environment sizing, optimizing the configuration, and understanding how much load an environment can handle.

How is customer service and technical support?

Initial technical support is excellent for common, easy to identify and medium complexity issues. For difficult, complex, issues it’s necessary to escalate things to get the experts from development and engineering involved to get traction on resolution. Otherwise, things can get stuck and drawn out.

How was the initial setup?

Initial setup is moderate to complex (there are a number of factors in the complexity of setup) and needs to be done with someone who has experience setting up and configuring CA APM environments.

Which other solutions did I evaluate?

No other products were evaluated prior to using CA APM.

What other advice do I have?

Having expertise from CA to help you with implementing CA APM is essential. It is also essential to understand that deep-dive instrumentation can have a significant impact on an application and even enabling or disabling certain agent features can cause significant impact. It is critical to understand how to configure the agent, what to configure, and test out a configuration before deploying to production. Also, any updates to applications must be tested with the agent in place, before deploying to production. If there are issues, it’s important to know how to start tuning down the level of instrumentation.

I will note that CA has, and continues, to bend over backwards in responding to our needs and requirements in ways that few vendors do.

Disclosure: IT Central Station contacted the reviewer to collect the review and to validate authenticity. The reviewer was referred by the vendor, but the review is not subject to editing or approval by the vendor.
1 Comment
User at CA, Inc.Vendor

Hello ITTechni9ef4,
Thank you for taking the time to provide your feedback!

I am glad to hear that your CA APM 10.2 stability issues have been addressed. Our Engineering and Support teams have worked hard to systemically address any stability issues that crop up, and they work hard to incorporate those improvements into new releases. Hopefully, as you continue to upgrade and try out new features, you will find that our previous work stabilizes your new trials.

Speaking of new things to try out, you are not alone in voicing concerns about our .NET agent. We have greatly increased our investment in .NET, and we plan to release a great new set of .NET agent features soon. If you get the chance, please join us as one of our customer design partners as we develop features, or try them out at release and provide your feedback. (New .NET features are in Beta today for design partners)

Assuming you get the chance to check out the new .NET agent, you will no doubt end up putting even more metrics on a dashboard. I appreciate your use of APM’s metric openness to use your expert knowledge to pull together the perfect dashboard for your teams. However, if you are like most customers we hear from, you will no doubt lament missing Event data in your dashboard. For example, listing out specific user events or error events in a table next to API performance metric time slices.

Momentum for addressing this Event shortcoming has gained speed as we continue to deepen our focus on user experience and user journey within AXA. AXA uses another open technology under the hood, Elastic, to store event data, and AXA comes with a product called APM Data Studio to visualize Event data.

This Event data is just part of the story, and we have been busy architecting the stitching of all our data into APM Data Studio via APMSQL. APMSQL gives you the power of SQL to sift through the mounds of APM timeslice data to easily find something like “the number of agents that saw response times over 5 seconds last week”. We believe that Elastic and APMSQL together will supercharge your dashboards. Should you decide AXA is not something you can embark on today, we do plan to offer APM Studio directly in CA APM.

Lastly, I’ll mention that we recognize the desire to strengthen APM’s self-healing assistance, but we are not yet comfortable with APM’s (industry wide) ability to correctly interpret application conditions. Triage and diagnosis are sensitive areas, and automation that improperly acts has serious financial impact. APM 10.5 is the first prescriptive APM that attempts to truly lead you through triage to actionable diagnosis in as little as five seconds. If you get the chance to check out 10.5, I hope you will agree that we are working hard to put real world first-of-its-kind smarts behind diagnosing application issues.

We call these diagnosing smarts Assisted Triage, and after we gather data on its performance in the field, we’ll have a much better sense of how close we are to accurately understanding application status. When we have the appropriate confidence that we can do more good than harm with auto-heal features, you will see us invest heavily in helping you take advantage of them.

Thanks again,
Nate Isley

31 January 17
Sign Up with Email