Implement reliable cloud-router SQLite backups
Implement and deploy an online backup solution for the cloud-router SQLite database on router.kaukus.no. Outcome: - Backups are transactionally consistent without stopping cloud-router. - Backups run automatically with bounded retention and failure visibility. - Backup storage is separate enough from the live database to cover database or deployment damage; off-host durability is addressed explicitly. - A documented, tested restore procedure verifies that produced backups are usable. - Existing deployment behavior remains intact. Current findings: - The live database uses WAL mode and passes SQLite quick_check. - The host has no separate backup filesystem or off-host backup client/configuration. - cloud-router was intentionally stopped and removed immediately before investigation; the data volume remains intact. Blocker: - Confirm the off-host object-storage destination and provision a dedicated bucket-scoped access key. A separate encryption key must also be stored outside the router host before the backup can be considered recoverable.
Checklist
0 of 3 complete- Deploy encrypted continuous off-host replication without stopping cloud-router
- Verify a real isolated restore and schedule recurring restore checks
- Add source-controlled production configuration and recovery runbook
- Area
- Streaming
- Project
- streaming-engines
- Kind
- Operations
- Priority
- High
- Task ID
tsk_gpsgynnaaa- Updated