Success rate per service in the last 24 hours. A green ring at 100% means every API call succeeded. Below 95% turns yellow (some calls failing), below 80% turns red (service is degraded). Click a card to filter errors below.
Failed API calls over time, stacked by severity. Red = hard errors (API returned an error response), orange = timeouts (API didn't respond in time), yellow = rate limited (hit quota/throttle), grey = aborted (request cancelled). A tall bar means many failures in that time bucket. Spikes suggest an outage or bug.
Average response time per service over the selected period. A spike means that service was slow during that window. Sustained high latency (e.g. Gemini >10s, Deepgram >5s) may indicate API issues or large payloads.
Individual failed API calls, newest first. Each row is one call that returned an error, timed out, was rate-limited, or was aborted. Use the service filter to narrow down. Hover over the message column to see the full error text.
| Time | Service | Operation | Status | Error Code | Message | Latency |
|---|
Automated root-cause analyses from the Gemini-powered error investigator. These run automatically when the app detects error patterns (3+ of the same error in 10 minutes). Red border = critical (needs immediate fix), yellow = warning (should investigate), blue = info (likely transient).