Operations field guide

A repeatable baseline for small infrastructure changes and incident response.

Use this list as a starting point, then attach commands and expected output for the specific service.

1. Record the current state and keep a restorable backup.
2. Check both the intended path and one independent path.
3. Validate DNS answers on more than one resolver.
4. Inspect the certificate presented over the network.
5. Verify real HTTP status codes, including a missing page.
6. Confirm firewall rules allow the expected caller and deny another.
7. Run application-level tests instead of stopping at TCP.
8. Review logs immediately after the change.
9. Reboot during a planned window when persistence matters.
10. Repeat the same checks before calling the work complete.

Minimum incident record

Capture the exact client error, server timestamps, affected routes, and the last known good state. Mark each statement as confirmed or hypothetical. Close the note with a mechanism, not just a description of the final configuration.

impact=which users and routes
observed=exact errors and timestamps
confirmed=tests and their output
cause=verified mechanism
prevention=monitoring or runbook update

Maintenance exit criteria

All required routes pass end-to-end checks, the website and error semantics remain correct, no unexpected firewall writers are active, and the backup can be located without searching through shell history.