A 15 Minute Incident Checklist for a Small Side Project

Recovery was not the only job

During a weekend database connection incident, I focused on restarting services and did not post a status update. Recovery was quick, but users spent that time seeing failures with no explanation. That was the gap I wanted to remove.

Minute 0 to 3: establish facts

Check error rate, last successful request, and affected features. Do not publish a guessed root cause.

journalctl -u myapp --since "15 minutes ago" --no-pager
curl -fsS https://example.com/api/health

Minute 3 to 5: say what is known

Post the impact and a time for the next update. A short message such as Login and saving may fail. Investigation is in progress; next update at 20:15. is better than silence.

Minute 5 to 15: use reversible mitigation

Restarting a worker, limiting traffic, or switching to read-only mode is safer than deleting data under pressure. Confirm backup status before any irreversible action.

Four lines after recovery

• Start and recovery times
• User impact
• Confirmed cause
• Prevention owner and due date

I fixed a connection return path before raising the database limit, then added external monitoring. The checklist is small, but it makes communication part of recovery rather than an afterthought.