AIOps: Observability + Machine Learning + Automate Everything You Can at DevOps Day Seattle
Come join me in {virtual} Seattle on May 4th for DevOps Days Seattle.
In my session “AIOps: Observability + Machine Learning + Automate Everything You Can”, I’ll be delving into one of the hot new topics in the DevOps world! What is “AIOps”? and why should you care? Let’s take a few minutes to explore those questions.
What is AIOps?
The “AI” refers to Artificial Intelligence – Using computers to accomplish tasks that normally require human intelligence. And of course, “Ops” refers to IT operational tasks. The mantra in the DevOps community is, “Automate everything you can.” And a lot of our focus is on automating routine, repetitive operational tasks so people can be freed up to focus on the work is best uses their abilities and intelligence.
AIOps is about expanding the boundaries of things that can be automated. After we have relegated the mind-numbing grunt-work to computers, it enables us to go even further – using automation to sift through the ever-increasing barrage of data to find what’s actionable so we can spend less time figuring things out and more time doing something about them.
The primary subset of AI that we can use for this is Machine Learning (ML) – Using statistical techniques and computational algorithms to recognize patterns in data. While the human brain is built for pattern-matching, the sheer volume of data in IT Operational environments is a significant limiting factor. ML thrives on the volumes of data that confound people, making it well-suited for this use.
Why should you care?
First, check the current marketing verbiage from the vendors who supply your monitoring and related tools. (I’ll wait a minute while you hit their websites … Did you?) If you did, I’m sure you saw mentions of AIOps. Every tool vendor claims they do AIOps. Of course, no single tool is the answer. But the tools you currently use (and the ones you’ll buy next month) are already part of the AIOps ecosystem, so you’ll need to learn to make good use of them.
Second, one of the most obvious uses of AIOps is diagnosing problems so they can be fixed quickly – arguably the most pressing (and vexing) issue in most IT shops. Image your tools detecting that there is a problem, sifting through tons of data to match the situation with prior events and notifying you of precisely what needs to be fixed before the users have even noticed it! Not only is this already within reach, but the longer you use these tools, the faster and more accurate they’ll become as the amount of historical data available to them grows.
Then there’s the interesting future possibilities! Like “self-healing” systems! Automation that goes beyond merely diagnosing and can take corrective action! Or automation that analyzes operational statistics and identifies ways to optimize performance or improve stability! The opportunities truly are endless.
If AIOps sounds interesting, then DevOps Days Seattle is a good place to start. And after attending my session, I’m sure you’ll find other valuable sessions too. I hope I’ll see you in {virtual} Seattle on May 4th!
