Everything's running smooth.

Submissions to MQD were taking 8–14 seconds instead of the usual 2–4 seconds for several hours on Apr 21. No claims were lost; all eventually succeeded. Root cause was an MQD-side endpoint upgrade we were not notified about. We've added a circuit-breaker that surfaces the slowness in the Claims Queue if it happens again.

A network configuration change at our Honolulu office disconnected the KHS hardware from the Tailscale tailnet for 14 minutes. The app showed loud network errors during this window; no fake-success fallback paths were triggered. DSWs in the field had their visits queued locally and synced automatically when connectivity returned.

Routine quarterly Tailscale key rotation took 6 minutes instead of the expected sub-30-second rotation. Mobile DSW clients with cached sessions kept working; only fresh sign-ins were affected. We've moved the rotation window to Sundays at 02:00 HST going forward.

A bad cache header on the /claims/queue route caused ~3% of requests to return 4xx for 8 minutes after a deploy. No data loss; users who refreshed got the correct response. We added a deploy-time validation step that catches this pattern.