Troubleshooting tree

Pick the symptom, then the capture. Change one thing.

1. No audio, agent shows Available Microphone permission, then UDP 3478, then VPN on the media path. Lab: webrtc.
2. ice_collection_timeout DNS for TurnNlb from that PC. Lab: webrtc.
3. Flow error and an empty Lambda log Association and the Lambda resource policy. AccessDenied never reaches your function. Lab: lambda-sim.
4. Lambda log is fine and the flow took Error Response shape or the 8-second wait. Lab: lambda-sim.
5. Caller hears silence and the contact ends An error branch is not connected. Read the flow log for that ContactId. Lab: flow-log-reader.
6. Wrong queue answers Priority and delay on the profile, not the queue name. Lab: routing-designer.
7. Recording missing Block enabled, bucket policy, key policy, then one new test call. Lab: find-fault.
8. CTR missing after a minute Wait. Five quiet minutes is the point to read Firehose errors. Lab: ctr-inspector.
9. Screen pop is the wrong person ANI matched a household. Pop the account id the flow collected. Lab: ecosystem-map.
10. WFM adherence is blank One agent event, a status map, and a published schedule. Lab: wfm-adherence.
11. 911 would reach headquarters The stored profile address is stale. Validate, store, test with 933. Lab: e911-checker.

Source: re:Post, flow errors