Notes from the on-call rotation

What we learn building an AI investigation kit for Kubernetes — including the failure modes agents miss, and the things we couldn't verify.

Engineering·30 July 2026·6 min read

I built a DevOps tool that actually reads your cluster

Every AI agent I'd used was asking me questions my own cluster could have answered instantly. So I gave it eyes — and then had to build a safety layer, because a live cluster connection is a loaded gun.

Read it