Every Jan. 1, as the New Year begins, we make resolutions. The idea is to take our new beginning and refresh ourselves with ideas to make “things” better — ideas to make us behave better, live healthier, help others or be a better Earth citizen. But resolutions don’t just have to be personal endeavors. Just as our individual resolutions can promote personal growth, IT organizations like DevOps (or Ops in general) can result in healthier applications — and we all know that healthier applications make happier customers. In the spirit of January 2020 (and my resolution to write more), here are three New Year’s Resolutions for application performance management (APM) operators (from Developers to Operations):
1: Dump Your Older Tooling Like the Back of Your Closet
Every January, I hit the closet — and I journey to the back. I’m looking for clothes that are out of style, clothes that don’t work with other items in my closet and — let’s face it — clothes that just don’t fit me anymore. I bring in some boxes (no plastic bags) and fill them up. I take out everything that doesn’t fit — wrong color, wrong size, and wrong style.
IT tools are no different. We outgrow them. Sometimes, we outgrow them because they just don’t have what it takes to go any longer. Other times, changes and digital transformation have made older tools less useful.
APM is a great example. Those cutting edge SOA monitoring tools (APM Generation 2) were designed 15 years ago (that’s NOT a typo). Before Facebook. Before the iPhone. In other words, before everyone essentially had a supercomputer in their hand. If you’re even thinking about running microservice applications, containers in your data center or embracing a multicloud environment, those old tools will eventually let you down.
It’s more than just supporting technology. Microservice applications don’t necessarily revolve around a central piece of Java — creating visibility gaps for traditional tools. And if you’re speeding up your deployments, you’ll need something with a little more responsibility than the minutes (or even hours) it takes to get feedback.
2: Embrace the Site Reliability Engineering (SRE) Concepts
Google vice president of engineering Ben Treynor is a very visible, very vocal advocate of Site Reliability Engineering as a way to make a difference in delivering performance-optimized applications to your end-users. The interview with Treynor over at Google’s Operations Blog is a great read — and invaluable for any IT Operations staff. At the heart of the interview is this answer to the question “what is Site Reliability Engineering?”
Treynor’s answer: “Fundamentally, it’s what happens when you ask a software engineer to design an operations function.”
Truthfully, you may not agree with everything Treynor says. That’s okay — operations, especially in mission-critical business application shops, are as unique to each company/organization as fingerprints. To me, though, a very important idea from the interview is the thought of taking a talented engineer — and have them apply engineering to IT Operations. It sounds so simple, but of course, we know it’s difficult to commit to that in real life.
At Instana, we transitioned our traditional IT Ops team to a focus on Site Reliability Engineering a few years ago. It made a huge difference in our ability to deliver the expected service levels — as well as solve problems quickly when they occur (using our own product to monitor of course).
Here’s the other great thing about Ben’s ideas — they drive unique solutions, but solutions that you already have the talent in house to develop.
3: Leverage Observability and Visibility
In case you don’t follow all the APM goings-on in the SocialVerse, especially Twitter, there’s an interesting debate going on between the old school APM providers and the new generation of monitoring tools. It can be boiled down to this: Observability vs. Visibility.
Traditional APM tools don’t worry about Observability — focusing on providing the most complete monitoring agent they can. Meanwhile, a new set of “application monitoring” tools have focused on UI, Dashboards and Intelligent Operations, and push Observability to avoid becoming victims of GIGO (Garbage In, Garbage Out).
It’s becoming somewhat of a religious debate, with both sides digging in. Well, do yourself a favor and avoid that holy war by embracing the best of both worlds. That means finding tools, especially APM tools, that support both the proprietary monitoring agent AND the nuanced context of developer-delivered observability through the growing popular Open Source monitoring technologies.
If you truly think about it, why settle for anything else. No matter which way you chose, you’d leave some extremely valuable information (metrics, etc.) on the table. By leveraging both methodologies, you guarantee yourself the best possible coverage — oh yeah, it will make the development team happier, too.
Okay — there you have it. Three resolutions for IT Ops (around APM). Feel free to throw on a few more of your own — less soda, laugh twice a day, write your own application proxy — you know — the easy stuff.
In all seriousness, when you stop growing personally, things can stagnate. That’s why resolutions are so important to individuals. Organizations are analogous to a living organism, too. An annual introspection of processes, goals and tools can make all the difference to being a high-performance team this year.
Feature image via Pixabay.