Why it happens
Requests or tokens per minute passed the plan's limit — usually a burst, or a loop, rather than steady growth.
How to tell
- 429 responses carrying a retry-after header.
- Bursts of calls inside the same second.
- A usage spike shortly before.
The fix
- Queue and back off: retry a 429 after the retry-after value, and cap concurrency.
- Trim prompts, and cache responses for repeated inputs.
- Ask the provider for a higher limit once the pattern is fixed, not before.
The prompt for your builder
Paste this into Lovable, Bolt, Cursor, Claude Code or whatever built the app. Replace anything in capitals. It never asks you to paste a secret into a chat; keys go into your host’s environment, by you.
Prompt for your builder
Add a small request queue with a concurrency limit and exponential backoff on 429 to every call my app makes to the AI provider, honouring the retry-after header, and cache responses for identical prompts for one hour.
How Keelnest catches it
Error capture groups the 429s under one fingerprint; Cost Guard shows the burst that caused them.
Know the moment this happens.
Keelnest watches your production app every five minutes and tells you in plain English what broke — with the fix. Free for your first app.