pgcopydb-operator

A Kubernetes operator that runs PostgreSQL migrations with pgcopydb. Declare a Migration resource to get a supervised bulk clone, optional logical-replication follow with controlled cutover, verification, and cleanup. Source and target are plain libpq endpoints, so it works with any PostgreSQL: managed, operator-run, or bare.

In development. One-shot clone, live migration with follow and controlled cutover, and verification are functional and e2e-tested. The API is v1beta1 and may still change. v1alpha1 stays served, and is deprecated.

Where to go:

• Quickstart: a first clone, end to end.
• Installation: the Helm chart and the CRD lifecycle.
• Configuration: clone tuning, filters, work volume, runner image, credentials.
• Live migration: follow mode, preflight, and the cutover runbook.
• Verification: pgcopydb compare after completion.
• Dry run, suspend, retries, deletion: checking a Migration before it runs, and its day-2 lifecycle.
• Monitoring: per-Migration metrics, the bundled Grafana dashboards, and alert rules.
• Migrating into CloudNativePG: CNPG targets and sources.
• Running the E2E suite against your databases: the product's own tests against a disposable pair of yours.
• Argo CD health checks: GitOps health for Migration resources.
• Troubleshooting: symptoms mapped to causes and fixes.
• Prerequisites: what your endpoints must provide; read this before a live migration.
• Planning checklist: decisions to settle before you create a Migration.
• CRD reference: every Migration field with defaults and validation.
• Conditions and reasons: the condition types and reason strings as API contract.