CommentaryIncident response
What MTTD misses
Separate detection from response and find delays that an improving average can conceal.
AI in production. Reliability in practice.
New here? Start with a reading path.Featured article
Timed-out automation writes: reconcile before retrying
For SREs and platform engineers building automation that can change production.
A lost response can turn a retry into a duplicate action. Learn what evidence you need before trying the write again.
Selected reading
Decisions worth getting right
Understand your detection gaps
Follow through after an incident
GuideIncident response
What did the postmortem fix prove?
Separate a completed task from evidence that the underlying risk was reduced.
Clarify who owns reliability
ExplainerPlatform engineering
The SRE and platform handoff
Clarify ownership across shared platforms, service reliability, and incident response.