Skip to content

CRD Reference

Retour uses six Custom Resource Definitions (CRDs) for declarative backup and restore operations. All CRDs use the API group retour.thebackup.company/v1.

Overview

CRD Purpose Scope
Backup One-off or on-demand backup of resources and PVs Namespaced
Restore Restore from a backup (full or selective) Namespaced
BackupSchedule Cron-based scheduled backups with retention Namespaced
StorageBackend Storage destination (S3, Azure Blob, GCS, encrypted) Namespaced
RestoreSync Warm standby: replay a shared backend's backups onto a second cluster Namespaced
BackupQuota Caps on what a namespace can consume Cluster

Installation

CRDs are installed with the Retour operator or manually:

kubectl apply -f operator/config/crd/bases/

Verify installation:

kubectl get crd | grep retour.thebackup.company

CRD Manifests

When you build the docs (task docs:build), CRD manifests are copied to the docs output for reference:

Typical Workflow

  1. Create a StorageBackend — Define where backups are stored (S3, MinIO, Ceph, etc.).
  2. Create a Backup — Point to a Deployment, StatefulSet, or Namespace; Retour discovers volumes automatically.
  3. Optional: Create a BackupSchedule — Automate backups with cron and retention.
  4. Restore when needed — Reference a backup by name or path; optionally restore specific resources only.