Operations

This section is for platform engineers and SREs deploying riverbank at scale. It assumes fluency with Kubernetes, Prometheus, and PostgreSQL.

Page	What it covers
Helm chart	Full values.yaml reference, upgrade, rollback
Multi-replica workers	Advisory locking, no duplicate work
Advisory locks	Lock keys, crash recovery, diagnostics
Circuit breakers	Per-provider config, states, recovery
Audit trail	What's logged, retention, querying
Backup and restore	pg_dump, point-in-time recovery
Secret management	Kubernetes secrets, Vault, rotation
Observability	Langfuse, OpenTelemetry, Prometheus, Perses
Scaling	Horizontal scaling, resource limits, bottlenecks
Upgrading	Alembic migrations, rollback, lock durations