How do you keep operational docs useful during an incident?

A shared documentation space can hold everything from runbooks to repository documentation, but it gets messy when nobody knows who owns which page. Give each document a clear owner so people know who can update it when the procedure changes.

Git is an option for keeping operational docs alongside the code. Markdown files in the repository can put service documentation and the service itself in the same place.

Docs are only part of the picture. Make dashboards and alerts part of the definition of done, then refine them as on-call work exposes gaps. Runbooks are useful too, but the monitoring needs attention as the service changes.

Rotating on-call work across developers and standardizing languages, frameworks, and databases between services can also reduce the number of different setups people need to support.