Operations runbook
This page does not replace your MongoDB Atlas or hosting provider console. It is a checklist for what to verify before and after launch.
Backups (MongoDB Atlas)
- Confirm Cloud Backup (or equivalent) is enabled for the production cluster.
- Confirm retention meets your recovery policy (often 7–30 days minimum for production).
- Perform a restore drill to a scratch database or namespace at least once per quarter. Document who ran it and the outcome.
Uptime monitoring
Configure external checks (e.g. Better Stack, UptimeRobot, Pingdom) for at least:
| URL | Purpose |
|---|---|
https://api.attunelogic.com/health | API liveness |
https://app.attunelogic.com | Web app |
https://attunelogic.com | Marketing site |
https://docs.attunelogic.com | Documentation |
Route alerts to a shared on-call channel (Slack or PagerDuty). For a small team, a dedicated operations phone is acceptable if someone acknowledges pages.
Sentry
- Configure alert rules for new issues in production and error-rate spikes.
- Keep DSN per environment; do not reuse production DSN in staging if you need clean signal.
Email deliverability
See Email deliverability for SPF, DKIM, and DMARC.