Triage The Latency Spike
Apply a structured troubleshooting sequence that prioritizes mitigation before deep diagnosis.
Live latency incident Checkout p99 is 8.4 seconds for ten minutes, traffic is normal, and the launch channel is asking for updates. A root-cause hunt that ignores mitigation lets user harm compound while the team gathers evidence. Google SRE troubleshooting Triage -> Examine -> Diagnose -> Test/Treat Use telemetry to narrow uncertainty, but protect users first when impact is active. Root-cause now Feels rigorous, but can leave users waiting. Users recover while evidence remains available for root cause. Mitigation and diagnosis are both engineering work. Sequence them by user risk. 01 Triage 02 Evidence 03 Treat First call Customers are…
Sign up free — one personalized lesson every day, matched to your role and goals.
Already have an account? Sign in