RCA

Your Incident Probably Did Not Have a Root Cause

Read enough postmortems and you notice they all end the same way. One box, one sentence, usually naming a deployment. Root cause: bad deploy. It is tidy, it closes the ticket, and it is almost always an incomplete account of what happened. Worse, it q…

The Alert Knows the Pod. The Pod Doesn't Know the Commit.

An alert fires. Checkout latency, p95, over threshold. What you have been handed is a pod name and a number. What you need is a function in a file in a repository, changed by a commit, shipped by a deployment, roughly forty minutes ago. Between th…

Load More
That is All