Follow a job and inspect its result

Use read_v2_job with the original workspace_id and application_id. Pass the submission's jobId as job_id and use the same authorized identity.

1. Check status on the durable job. For an event history, call read_v2_events in the same application. Start at after_sequence: 0 with a bounded limit. Continue with nextSequence as after_sequence and match events by job_id.
2. When the job reports published, record its receipt_revision. Repeat the same bounded facts.cube.scan used before the write.
3. Check the same coordinate for 12.50000000 and a new numericChangeId. In this isolated example the row's lastRevision equals the job's receipt_revision.

For failure, cancellation, or inspection-required state, read the exact diagnostic and determine whether any effect committed before considering a retry. See Monitoring and recovery tools.