Traffic Routing Incident Response Playbook

Determine whether traffic receives no response, the wrong destination, a redirect loop, stale content, missing attribution, or an unusual decision ratio. Identify affected hosts, paths, sources, regions, and start time before editing configuration.

Pause an affected flow, select a known fallback, or isolate a broken target according to the runbook. Avoid global filter changes when the incident is confined to one campaign or delivery mode.

Save representative event identifiers, configuration state, destination health, DNS observations, and recent change history. Evidence collected before a fix helps distinguish the trigger from side effects introduced during response.

Check host and path matching, request signals, rule precedence, risk decision, selected destination, and delivery response in sequence. Stop at the first divergence from the expected fixture rather than tuning later layers.

Replay affected and neighboring acceptance cases, monitor the repaired scope, then document root cause and prevention. Remove temporary containment only after ownership and rollback criteria are explicit.