CALIBER
Quickstart
Operate CALIBER

Backup and recovery

The operator view of what CALIBER state must be preserved, what can be rebuilt, and where disaster recovery stops and asset-specific rollback begins.

OperatorRunbookGA
PrerequisitesA CALIBER deployment or disaster-recovery review
Reviewed 2026-08-10 · current main branch docs contract

Use this page when you need the operator view of what CALIBER state must be preserved, what can be rebuilt, and where the disaster-recovery boundary stops and asset-specific rollback begins.

At a glance

SurfacePreserve it?Why it matters
Metadata databaseyescontrol-plane records, governed state, releases, jobs, and references
Object store datayesuploads, previews, extracted artifacts, and log sink material
Workflow storageyesrun workspaces and file-backed workflow evidence
Generated docs/UI assetsrebuildablethese come from repository sources and build outputs

1. Back up the state that cannot be re-derived safely

The safest minimum backup set is:

  • metadata database
  • object-store content
  • workflow storage

Without those together, a restore may recover only part of the product state.

2. Distinguish disaster recovery from feature rollback

Recovery of the deployment itself is not the same as rolling back a governed asset release. Use the disaster-recovery plan to restore the platform, then use asset-specific rollback or release controls where needed.

3. Verify restores, not just backups

The useful operator question is not “did a backup job run?” but “can this deployment restore to a usable, trustworthy state?”

After a restore rehearsal, verify:

  • readiness
  • storage access
  • release and review history visibility

CALIBER : Contextual Adaptive Lifecycle for Intelligent Build, Evaluation, and Refinement — this page is generated from the authoritative Markdown sources in docs/ and the repository-level ARCHITECTURE.md.