What this agent helps you do
A Cloudflare incident response agent helps engineers collect relevant platform context during an outage or degradation. It can summarize what changed, what signals are visible, and which next checks are most important.
When to use this workflow
Use it during edge errors, worker failures, unusual traffic, performance regressions, DNS or routing questions, or after an incident when preparing a postmortem.
How Cloudflare gives the agent context
Connect the plugin and constrain the scope to the zone, worker, deployment, or timeframe involved. The agent should gather available configuration and observability context, then distinguish verified facts from hypotheses.
Example starter prompt
Investigate this Cloudflare-related incident for the last two hours. Summarize visible symptoms, recent changes, likely causes, recommended checks, and any actions that require approval before changing production configuration.Suggested workflow steps
Define the incident window, collect Cloudflare context, compare recent changes with symptoms, and rank hypotheses. The agent should include rollback or mitigation ideas only as approval-ready recommendations.
Expected handoff
The output should include timeline notes, evidence, likely causes, unanswered questions, and safe next steps. Pair it with GitHub or Sentry when code changes or application errors may explain the incident.