Database operations¶
OpenIdentityStack uses PostgreSQL for application data and persisted data-protection keys.
Migration model¶
Schema updates are handled by the dedicated DbMigrator rather than by the API process itself. In the bundled Kubernetes deployment, the migrator runs as a job before the API rollout is considered complete.
Backup and restore expectations¶
Your database runbook should include:
- regular backups
- restore testing
- a plan for data-protection continuity
- rollback expectations when an application release is reverted
First-run seeding¶
The current deployment paths support two different seed shapes:
- development seed data for local AppHost flows
- optional production admin seeding through
open-identity-stack-admin-seed
Treat production seeding as a bootstrap event, not a permanent configuration source.
Checks after a restore¶
After restoring a database:
- verify the API can start
- verify
/health - verify sign-in works
- verify at least one client application can obtain a token
- verify data-protection-dependent flows still behave correctly