Skip to main content

Logs, recovery, and common failures

Hook and worker logs contain operational metadata and must remain owner-only. The managed log-rotation definition and rendered configuration are part of doctor evidence. Certification proves the service definition is installed, loaded, rotates the configured file, and preserves mode 0600.

Recovery order​

  1. Stop new submissions at the caller boundary.
  2. Preserve queue records, operation receipts/events, snapshots, plugin pointers, and logs.
  3. Check /health and /ready separately.
  4. Reconcile submitting and accepted memory records by operation ID.
  5. Restore worker service state and let persisted plans resume.
  6. Verify current snapshot pointers and active plugin generation.
  7. Run read-only doctors, then the mutating receipt certification if authorized.

Common failures​

FailureMeaningRecovery
/health 200, /ready 503Process live; ledger unavailableRepair storage/coordinator before writes.
Receipt remains non-terminalWork is blocked, interrupted, or waitingInspect blocked_by, executor fields, and ordered events.
Worker service loaded but queue growsWorker cannot claim or reconcileInspect queue ownership, logs, binary signature, and Memory readiness.
Plugin verify finds stale pathHook/target bypasses currentRestore stable dispatcher or verified target receipt.
Snapshot pointer invalidAtomic publication did not complete or path was alteredRepoint to a complete generation or republish.
Doctor warningOptional/degraded surfaceRecord explicit release disposition; never relabel as green.
Exec health is live but readiness failsSidecar bound before a sandbox, trust, ledger, or CAS subsystem became readyRead the named readiness subsystem and run the non-mutating execution doctor.
Exec receipt exists but response was lostTerminal evidence committed before the caller received itRead status/events by run ID; do not resubmit with a different request ID.

Do not erase recovery records to make a doctor green. Repair the active pointer or service state while preserving published generations, terminal receipts, and command evidence.