CloudFront recovery console

First-class governed Amazon CloudFront recovery-action lifecycle. The Detect → Investigate → Decide → Approve → Act → Verify → Learn flow lives at/cloudfront-recovery/<incidentId>; the test-fixture button below seeds a per-user synthetic incident in supported development/test environments only.

Fresh-run preparation

Development/no-spend only. This captures real player-SDK evidence and approvals; it never authorizes or runs CloudFront.

provider unexecuted · no external success

Fresh run version *Development target *Named target evidence source *Cohort scope *Content scopeContent version *Bounded path scope *Observation window start *Observation window end *Restoration check *Cleanup check *

Approved playback criteria

Decoded playback minimum *Fresh content minimum *Stall-rate maximum *Player-error maximum *Latency maximum (ms) *

Governed approval for this exact run

Approval is attributable to a distinct actor and covers only the saved target, paths, cohort, content version, and observation window. It never enables provider execution.

Cleanup and restoration verification

Restoration evidence source *Independent observation time *Cleanup summary *

No-spend preparation harness

Creates synthetic preparation evidence only. It never calls AWS; recovery can only be verified by run-scoped player SDK observations.

provider action unexecuted

Loading active workspace…

Idempotent — re-running converges. Available only in supported development/test environments with ENABLE_TEST_FIXTURES=1.