Background jobs / Decision guide
Dead-Letter Queue Triage
Should this failed job be retried, repaired, quarantined, or investigated? Start with the evidence, not the retry button.
Interactive calculations run in your browser; the initial example is pre-rendered. There are no accounts, uploads, or live queue connections. Inputs stay in page memory; the site does not persist them, put them in URLs, or send them to analytics. A worksheet download includes only what you explicitly export.
Suggested disposition / human review required
Reconcile before considering a retry
Who should own the decision: Application owner and downstream service owner.
- Locate the original operation key and downstream receipt.
- Resolve whether the effect happened; preserve an UNKNOWN state if evidence is unavailable.
- Prevent competing replay attempts while investigation is in progress.
Evidence to retain in your own approved system
- Failure stage, redacted error category, and handler/schema versions.
- Original intent identity, downstream receipt, and the evidence supporting the effect state.
- Job age, applicable retention window, current authority, and the decision owner.
- Approved disposition, selected scope, stop conditions, and unresolved questions.
For a replay candidate, complete the Replay Readiness Checklist before any operational action.
Assumptions and limits
- This guide evaluates your reported evidence; it cannot independently diagnose a job or certify replay safety.
- No job, queue, payload, or log is uploaded. No retry, delete, or requeue action is performed.
- Discard only under an approved retention/business policy, with an audit record and the right owner.
- Queue-specific redrive, ordering, deduplication, and retention semantics still apply.