-
Runbooks are a feature, not a chore
The team that writes the runbook before the incident is the team that sleeps through the next one. A short argument for treating operational docs like product work.
-
Why we stopped patching servers
Swapping the whole instance is less work than keeping one alive. How immutable infrastructure quietly removed an entire class of Friday-afternoon surprises.
-
Alert on symptoms, page on pain
Most alerts fire on things nobody should be woken up for. A simple rule for deciding what deserves a page at 3am.