fce43b80

docs: correct Attic postmortem monitoring findings (#288)

## Correction

The original postmortem said Attic had no health check. Gatus already checked it every minute and worked:

- first failure: 10:10:57 CDT
- ntfy alert after three failures: 10:12:58
- recovery alert: 11:21:58

The response failed because one noisy monitoring stream made the important Attic alert look routine. Buildbot and voicemail also fired transient alerts at nearly the same time.

## Changes

- correct duration and timeline
- replace the false missing-monitor claim with alert fatigue
- mark the DB URL audit and fixed endpoint checks complete
- add bounded follow-ups to rank noisy checks and split alert urgency
- state why Attic should not trigger automatic rollback

Docs only. All pre-commit hooks pass.

Co-authored-by: Ammar Nanjiani <ammar.nanjiani@gmail.com>
Reviewed-on: https://codeberg.org/ananjiani/infra/pulls/288

Total

2h34m23s

Timeline