Aglet

Verify a durable fix for a recurring release regression

Verification should test durability, not only immediate quiet. Recreate the condition that failed before, cover the scope that was missed, and compare behavior after the correction with both prior episodes and a known-good control.

Check whether the outcome improved

  1. Replay the failure boundary

    Use the same request, job, data shape, environment, cohort, and release interaction that defined recurrence. Record any difference that prevents a faithful comparison and adjust confidence accordingly.

  2. Cover the missed condition

    Include the path, version pair, configuration, client, or delayed process identified as absent from the earlier verification. Review successful completion, errors, retries, latency, and downstream effects.

  3. Watch for return

    Observe through the recurrence window and compare with prior episodes under comparable traffic. Apply the durability rule; if the window is incomplete, leave the regression open with a dated follow-up.

What to carry forward

Verification should say durable recovery, partial coverage, recurrence, or inconclusive, backed by the original and missed conditions. Close only when the returning symptom has had a full comparable observation window. Exercise the original failing path and a nearby unaffected path under the current release, then compare with the earlier fixture or observation before calling the recurrence contained.

Keep the decision with the work.

Use a Work Item in Aglet to record the problem, the evidence you have, and the next decision. Add an owner and priority, then keep updates in the discussion so the next person can follow the reasoning.

Create an account See the product workflow