API operations
Traffic, reliability and the work behind each lookup.
Checking owner access…
Request volume
Daily recorded requests and server errors.
View daily data
| Date (UTC) | Accounts ≈ | Requests | Credits charged | 5xx errors | Known cost estimate |
|---|
Response time
Daily latency bucket upper bounds, in milliseconds.
Bucket estimates, not exact percentiles. Overflow above 30s has no upper bound and is shown in inspection rather than plotted.
Provider calls
Requests to upstream services, ranked by volume.
View provider data
| Provider | Calls | Failures | Input tokens | Known cost estimate | Unknown-cost calls |
|---|
Cache performance
Share of recorded cache hits by lookup stage.
View cache data
Hits include shared work in progress. Calls avoided are estimates from provider calls observed when an entry was filled, not measured savings. Failed fills are excluded.
| Cache | Hits | Fills | Shared results | Est. calls avoided | Cache errors |
|---|
Latency distribution
How many requests landed in each response-time bucket.
Recent requests
The latest 50 retained requests, across all dates.
| Time | Account | Query / result | Method | Status | Latency | Credits | Details |
|---|
No requests match these filters.
Request details
Data sources and limitations
Operational counters are retained for 31 days. Charts leave days without recorded counters blank; a gap does not imply uptime or zero activity. The current UTC day is partial. Approximate active accounts include operator and test accounts and are not customer traction.
Query records expire after 30 days; request metadata after 31. Use the business overview for durable customer metrics and historical coverage.