Release quality playbooks
Connect release context with observed behavior and follow-up work. Choose a scenario, then the decision you need to make.
Post-release error spike
An error signal rose after a release, but timing alone does not establish that the release caused it.
Partial rollout divergence
A staged release behaves differently across cohorts or regions, leaving the overall deployment result misleading.
Environment mismatch
A release or signal appears in one environment while the expected version, configuration, or evidence belongs to another.
Missing commit context
A deployment or regression lacks a trustworthy commit, branch, or repository reference, slowing comparison and review.
Stalled rollout
A rollout remains in progress longer than expected, but the recorded state does not show whether progress or exposure continues.
Delayed symptom onset
A release is followed by a symptom only after a delay, making the relevant change window easy to misread.
Version skew
Different services, workers, or environments run incompatible versions, producing behavior that aggregate release status hides.
Configuration drift
Runtime configuration differs from the reviewed baseline, making a release outcome hard to attribute or reproduce.
Rollback uncertainty
A rollback is discussed or recorded, but the evidence does not establish which version serves traffic or whether symptoms changed.
Release regression recurrence
A previously resolved release regression returns, requiring comparison of prior fixes, affected paths, and changed conditions.