Troubleshooting

Everyone appears to have the same address

Check core trusted-proxy resolution. Sentinel deliberately does not trust arbitrary forwarding headers. Return to Monitor and correct the core integration before enforcing address limits.

Visitors are locked out

Use File Manager recovery, correct rules/provider configuration and test an ordinary account before removing the flag.

No events

Check enabled/licensed state, request-hook heartbeat, exemptions, mode, retention and the kind of event you expect. Not every normal request creates a flagged event.

Queued provider/catalogue/webhook work never completes

Check the addon scheduler under the correct hosting runtime and the job state. A queued refresh does not mean network acquisition succeeded. Test the real provider/TLS setup privately.

Redis probe succeeds but behaviour is unproven

Probe evidence is narrower than real outage, eviction, restart, row-lock race and load testing. Keep primary off until those gates pass on the actual service.

An older README contradicts a newer feature

Use CURRENT-COMPLETION.md and SETUP-beta75.md/beta73.md for current state. Historical sections saying Redis primary or a provider feature is absent describe earlier releases.