Delivery outcomes from the latest seven UTC dates, when a customer delivery sample exists.
Successful deliveries
0
Failed deliveries
0
Last sampled success
No recent sample
Methodology
What these numbers do—and do not—prove
API reliability. Rolling 24-hour completed 2xx and 5xx customer-call sample. Intentional validation, authentication, billing, and rate-limit outcomes plus owner and shared-Demo traffic are excluded.
Latency. p50 and p95 are service-side latency for successful calls in the same 24-hour sample. The live probe is a web-to-API health round trip and is not an end-user SLA.
Data freshness. Newest created_at timestamp returned by the fixed public Search Demo query. It proves that sample, not completeness or a platform-wide freshness guarantee.
Monitors. Successful and failed deliveries aggregated over the latest seven UTC dates. No recent sample is reported explicitly instead of inferring delivery latency.