Disaster recovery: backups + restore runbook

[MrImmortal09]

Context

If the DB is wiped or corrupted, we need to restore within a day without losing more than a few hours of data.

Acceptance criteria

• Automated daily MongoDB backups — use MongoDB Atlas continuous backup (if on Atlas) or scheduled mongodump to R2/S3
• Retention: 7 daily + 4 weekly snapshots
• R2 bucket also versioned for 30 days
• docs/runbooks/restore.md with step-by-step recovery using mongorestore or Atlas point-in-time restore
• Annual restore drill scheduled

Depends on #1, #3.

[Abhishek-Dige]

Hi! I’m planning to take this issue.

Since the current setup seems to use local MongoDB (no Atlas/S3/R2 yet), I wanted to confirm the expected scope:

Would you prefer:

1. A working local backup system (mongodump + local retention) with optional S3/R2 integration via environment variables, or
2. A full production-oriented setup assuming Atlas/S3/R2 (with placeholders/config docs)?

My plan is to:

• implement local backups for dev/testing
• design scripts to support S3-compatible storage (S3/R2) via env config
• add a detailed restore runbook covering both local and production scenarios

Let me know which one do you prefer.

[MrImmortal09]

Let's connect to Cloudflare R2 first and then have this issue solved.