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.

​
Control-plane backup and disaster recovery

Kindling splits durable state into two operator concerns:
  1. Control-plane state — PostgreSQL rows: organizations, projects, deployments, routing, cluster_settings, cluster_secrets, volume metadata, audit tables, and more.
  2. Workload data on persistent volumes — guest filesystem data backed by Kindling volume cold backup / restore to object storage.
Volume backup does not replace Postgres backup. If you only snapshot volumes, you can still lose the entire control plane.

​
Canonical contract

The full technical contract (RPO/RTO planning, recovery ordering, host-local state) lives in the repository: Supporting context:

​
Quick checklist

  • Postgres backups (or equivalent) cover the Kindling database used by every kindling serve process.
  • Every host has identical /etc/kindling/postgres.dsn (and postgres.replication.dsn when using PgBouncer) documented for restore.
  • Volume backup credentials and policies are documented separately from DB backup (S3-compatible settings in cluster secrets).
  • You know where ~/.kindling/server-id and WireGuard keys live on each node if you rebuild hardware.

​
Procedures

Step-by-step backup and restore validation lives in Backup and restore.