BrainBlog for Evolven by Jason Bloomberg
Managing incidents in any operational environment, no matter the complexity always boils down to this simple, three-step process:
Identify a problem ? Figure out what caused the problem ? Fix the problem
This reactive process is so ingrained in the way operators think about incidents that entire product categories have grown around it, from traditional IT incident management to AIOps to observability.
Nevertheless, this ‘see a problem, fix the problem’ approach is fraught with challenges. The good news is that there’s a better way: start with the causes and predict the effects – in other words, take a proactive approach to incident management.
Problems with Reactive Incident Management
To identify problems, operators first look to observability tooling. Observability provides telemetry in the form of logs, traces, and metrics – vital information about the behavior of various systems and applications.
If there’s a problem, it should turn up in these observability data. In other words, observability provides insight into the effects or symptoms of a problem, not its causes.
Click here to read the entire article.
