Track · 2:54 · Liner note
The API Timeout
An API timeout happens when a request to a remote service takes longer than the client is willing to wait, so the call is aborted without a result.
The request went out, nothing came back, and after thirty seconds your client gave up. The API timeout is the most ordinary failure in data ingestion and the one most often handled badly: no retry, or retries that hammer a service already struggling.
Do it properly: set explicit timeouts, retry with exponential backoff and jitter, make the requests idempotent and record what was missed so a backfill can close the gap. You cannot control the partner’s uptime, only how gracefully you fail.