Practice: diagnose a failed run and rerun only its failures

Use a training job with a failed execution. Choose its build ID from the search results and use that same ID for details, logs and the rerun.

Read the first failing keyword and its error in ./logs/failed-run.log. Decide whether the cause is test data, access, a locator or application behavior. Correct the cause; if you changed an uploaded script, use the file replacement workflow in 10b before retrying.

$ agentia testing job runs search --project <project-id> --job <job-id> --status failed$ agentia testing job runs get <build-id> --project <project-id> --job <job-id>$ agentia testing job runs logs <build-id> --project <project-id> --job <job-id> \  --output ./logs/failed-run.log# Resolve the cause first; upload the corrected script if needed.$ agentia testing job run <job-id> --project <project-id> \  --rerun-failed <build-id> --timeout 15 \  --save-artifacts ./artifacts/retry.zip --xunit ./reports/retry.xml

Check your results

• The rerun targets failures from the selected build, not an unrelated execution.
• Inspect retry.zip and retry.xml after completion; compare the failed test names with the original run and confirm their new results.
• If failures remain, keep both sets of evidence and investigate further. After a successful focused rerun, run the full suite to check for wider regressions.