Skip to main content

backup status

Reports CNPG (PostgreSQL) and Velero (volume/PVC) backup health, as evaluated by the in-cluster backup-exporter. Read-only: it changes nothing on the cluster or in Git.

Usage​

kubeaid-cli backup status          # human-readable report
kubeaid-cli backup status -o json  # the exporter's response, verbatim
  • Uses your current kubeconfig, like kubectl. If kubectl get svc -n monitoring works, so does this.
  • Finds the exporter itself — monitoring first, then every namespace. Nothing to configure, no general.yaml involved.
  • Reaches it by port-forward, exactly as kubectl port-forward would, so it works through netbird's ClusterProxy. Your identity needs what kubectl port-forward needs in the exporter's namespace.
  • Always exits 0 when it can report, however unhealthy the backups are. Non-zero means it could not fetch.

Output​

collected 2h 57m ago: cnpg | velero

Operator errors:
  cnpg: s3_list_failed

NAMESPACE    RESOURCE       TYPE           STREAM    METHOD            LATEST AGE   STATUS
demo         demo-pgsql     cnpg_cluster   logical   CronJob           none         collector_error (cronjob_not_found)
demo         demo-pgsql     cnpg_cluster   wal       Barman            4h 9m        healthy
demo         demo-uploads   pvc            volume    PodVolumeBackup   14h 57m      healthy
monitoring   obs-postgres   cnpg_cluster   logical   cronjob           3d 2h        exceeds_rpo
monitoring   prom-data      pvc            volume    CSISnapshot       14h 57m      healthy
4 resources: 2 healthy, 1 exceeds_rpo, 1 collector_error
  • Freshness line — how long ago the exporter last looked. One age when the collectors agree, per-collector when they don't (collected: velero 5m ago | cnpg never).
  • Operator errors — failures with no single resource to blame, e.g. Velero cannot list its bucket. Everything below them is that much less trustworthy.
  • Rows — one per backup stream, sorted by namespace then resource. A CNPG cluster gives a logical and a wal row.
  • STREAM / METHOD — what is backed up, and what takes it. METHOD names the thing to go look at when a row fails: CronJob for CNPG's logical dump, Barman for its WAL archive (the barman-cloud plugin in the Cluster CR), and Velero's own PodVolumeBackup / CSISnapshot.
  • LATEST AGE — age of the newest backup (2h 57m, 3d 2h), none when there is none at all, - when the exporter published nothing to measure.
  • STATUS — healthy, exceeds_rpo, no_backup, collector_error (reason in parentheses) or unknown.
  • Summary — counts resources, not rows. A resource whose streams disagree is counted once, under its worst one, so the figures never add up to more than you have.
  • Every row carries its namespace, so backup status | grep -v healthy gives you complete lines.

When it cannot report​

SituationWhat you see
backup-exporter not installed"no backup-exporter service found ... is the backup-exporter chart installed?"
Its pod is not Running + Ready"no ready pod behind service <namespace>/<name>"
Two exporters, neither in monitoringEvery namespace/name found, rather than a guess at which to use
The connection never comes upA timeout after 15s, naming the pod