Skip to main content

Documentation Index

Fetch the complete documentation index at: https://docs.kindling.systems/llms.txt

Use this file to discover all available pages before exploring further.

​
Backup and restore

This runbook covers two separate backup surfaces. Treat them as complementary.

​
1. Control-plane (PostgreSQL)

The Kindling API and reconcilers store authoritative state in PostgreSQL. Recovery requires a consistent restore of that database (plus valid DSN files on each host). Operator actions (outline):
  1. Follow your organization’s Postgres backup tool (pgBackRest, Barman, cloud snapshots, etc.) with a cadence that meets your RPO.
  2. Ensure restores preserve logical replication compatibility documented in High availability (wal_level = logical, listener expectations).
  3. After restore, verify DSN files on every Kindling host still target the correct primary or pooler entrypoint.
  4. Restart kindling serve units in your topology; run the validation checklist in Production Setup.
  5. Inspect cluster audit events after recovery if you performed admin operations during the incident (see docs/cluster-audit-events.md).
For semantics and ordering, read docs/control-plane-backup-and-dr.md.

​
2. Persistent volumes (application data)

Kindling supports cold backup and restore of project volumes to S3-compatible storage when the cluster is configured with volume backup settings and Cloud Hypervisor workers. Operator actions (outline):
  1. Configure cluster-wide volume backup credentials (object store bucket, keys) via the product’s settings / secrets surface.
  2. Use dashboard or API to queue a backup for a project volume when needed.
  3. To recover data, use restore to a selected worker per product docs; validate the workload afterward.
Note: Scheduled automated volume backups are roadmap; policy fields may exist before automation ships—see Secrets and volumes.

​
3. Validation after any restore

  • GET /api/meta succeeds (platform admin).
  • GET /api/servers lists expected nodes; heartbeats healthy.
  • Deploy or redeploy a canary project; confirm edge routes to workloads.
  • For volume restore: confirm application-level data integrity inside the guest.