What causes a 429 Too Many Requests error?

429 is triggered when a client exceeds the server's rate limit — too many requests per time window (per API key, IP, or user), daily/monthly quota exceeded, or the server is throttling under heavy load.

How do I implement retry logic for 429?

Use exponential backoff: wait 1s, then 2s, 4s, 8s between retries. If the response includes a Retry-After header, wait at least that long. Cap maximum retries (e.g. 5 attempts) and maximum wait time. Never retry immediately.

What are the standard rate limit headers?

X-RateLimit-Limit (max requests per window), X-RateLimit-Remaining (requests left), X-RateLimit-Reset (when the window resets, Unix timestamp or seconds), and Retry-After (how long to wait before retrying).

How should I design rate limits for my API?

Choose limits per API key, user, or IP. Use sliding window or token bucket algorithms. Document limits clearly. Return 429 with helpful headers. Consider different tiers for different API plans. Always rate limit by the most specific identifier.

Do search engines respect 429?

Yes. Well-behaved crawlers (Googlebot, Bingbot) respect 429 and Retry-After, slowing their crawl rate. Use 429 (not 503) when you want crawlers to slow down but keep trying. If crawl budget is an issue, also check your robots.txt and crawl rate settings.

429 Too Many Requests: Rate Limits and Client Strategies