Reliability and error handling
Rate limits, 429 and Retry-After
Bound resource use and guide clients.
Understand the concept
Limits can use account, IP, API key or operation cost. IP-only policies can penalise shared networks; trusting forged forwarding fields undermines counters. Login, uploads and AI requests have different costs.
429 communicates limiting. Retry-After can be seconds or an HTTP-date, so parse deliberately. Bound retries with backoff, jitter and a total deadline.
Connections and large bodies can consume resources independently of request count. Combine limits with bounded queues, concurrency controls and monitoring. Example limits are illustrative, not production capacity recommendations.
Protect expensive AI and contact-form endpoints.
Immediately retrying a 429 indefinitely.
One topic. Three levels of depth.
HTTP/1.1 429 Too Many Requests
Retry-After: 30An advisory thirty-second delay.
This page does not send requests to external servers. Messages are readable HTTP/1.1 illustrations, not HTTP/2 or HTTP/3 wire dumps. example.test is reserved for examples. Run commands and JavaScript only in your own test environment; never publish real tokens or cookies.
Try it yourself first.
Is Retry-After always an integer?
Show the solution and explanation
delay-seconds OR HTTP-dateNo; it can be an HTTP-date.