The API client log records every call made by registered API clients — what they asked for, when, how the server responded, how long it took. It's the primary tool for debugging integrations, capacity planning, and abuse detection.
What Each Row Includes
- Timestamp with millisecond precision.
- Client identity — which registered API client made the call.
- HTTP method and path — e.g.
POST /api/v1/records/orders. - Request headers and (truncated) body — for diagnosis.
- Response status — 2xx success, 4xx client error, 5xx server error.
- Response duration in milliseconds.
- Bytes sent and received.
- Source IP and user agent.
- Rate limit headers at the time of the call.
Filtering and Search
- By client — drill into one integration.
- By status code class — 4xx for client errors, 5xx for server errors.
- By endpoint — see which paths are hottest.
- By duration — find slow requests.
- By date range.
Worked Examples
- Integration broken: Customer reports their CRM sync stopped. Filter by their API client, last hour, status ≥ 400. See 401 Unauthorized — their token has been rotated and the integration wasn't updated.
- Capacity planning: Group by endpoint, sum of calls last 30 days. The most-called endpoint accounts for 60% of API traffic — worth optimising.
- Suspicious activity: Sort by error rate descending — a previously-quiet API key suddenly has 80% 4xx rate; might be misconfigured or compromised.
- Slow endpoint: Sort by duration descending. The 99th-percentile is 8 seconds; investigate that endpoint.