Operate what exists today.

Orbita does not yet ship a complete production operator kit. These procedures support evaluation and upgrade testing; dashboards, alert thresholds, backup/restore, and common incident runbooks are roadmap work.

Check the cluster

orbita cluster describe
orbita cluster ping
orbita cluster ready

Use --output json in automation.

Rolling upgrade

Follow docs/UPGRADES.md. Readiness gates the rollout, and finalizing a cluster version is a one-way operation. Restore tooling does not yet exist, so do not finalize an evaluation cluster you may need to downgrade.

Object-store CI incidents

The repository includes a focused live object-store CI runbook  for credentials, AWS/R2 checks, and Terraform state. It is not a general production-cluster runbook.

Missing runbooks

  • Backup and point-in-time restore
  • Node data-directory loss
  • Quota saturation and capacity thresholds
  • WAL lag and failover alerts
  • Security credential and certificate rotation

Track this work on the roadmap.