OpenAI API field note
OpenAI 429: insufficient_quota vs rate_limit_exceeded
Read the OpenAI 429 error code before applying backoff: account quota and a temporary request or token rate limit need different fixes.
The HTTP status alone is ambiguous. A temporary request or token limit can recover as the usage window resets; insufficient_quota points to project billing or limits and does not become healthy because a client sleeps longer.
Treat the complete response as an evidence record: status, headers, provider code, request identifier, method, and raw body. The sequence below separates what the response proves from the checks still needed before a safe retry or code change.
Diagnostic procedure
Work from evidence to recovery.
- 01
Read error.type and error.code
Do not branch only on status 429. Preserve both fields and the provider request identifier.
- 02
Match the owning project
For quota errors, confirm the API key’s project, billing state, budget, and organization or project limits.
- 03
Pace temporary limits
For request or token windows, reduce concurrency and batch size, then apply capped jittered backoff.
429 {"error":{"code":"insufficient_quota"}} → retry loopinsufficient_quota → billing/project limits rate_limit_exceeded → pace + bounded backoff
- 100
- window limit
- 0
- remaining
- 17s
- Retry-After
- 0
- safe req/s
Precedence Honor Retry-After when supplied. Treat reset and remaining fields as pacing evidence, not permission to burst.
Clock form X-RateLimit-Reset is often epoch seconds; the standardized RateLimit-Reset field may represent delta-seconds depending on the provider implementation.
No account · no upload · no endpoint calledOpen the full rate-limit analyzer and operating notes →
FAQ
Before you ship
Does APITC send this evidence to an API?
No. The matching and calculations in the linked workbench run in the active browser tab.
Should every OpenAI 429 response be retried?
No. Retry behavior depends on the method, idempotency protection, provider instructions, and whether the failure is temporary.
Protocol behavior checked against the OpenAI API error codes. Recheck your pinned provider/API version before production deployment.