How-to guideApplies to: Unified engine deployments
Inspect an existing workload execution without starting replacement work. Enter its workload ID and run ID under Example values.
Investigate a workload run
Set an existing workload ID and run ID. These requests inspect execution and notification delivery without starting another run.
Replace the placeholders in your client. Keep returned IDs for later requests. Paths below use the native API prefix; adjust it for your deployment.
1. Read the run
Endpoint reference.
2. Read tasks
Endpoint reference.
3. Read delivery history
Endpoint reference.
Read the recorded invocation and state. Compare revision, parameters, data window, queue reasons, and task attempts. A current editor definition can differ from the version recorded by the run.
Notification history is separate from data execution. An exhausted delivery does not turn a successful workload into a failed one. Use delivery retry to resend an event; use selective rerun to repeat eligible tasks.
For a write with an uncertain outcome, keep the original identities and inspect commit evidence. Do not create a fresh append because the browser lost its response.Last modified on November 4, 2026
