## TL;DR

Add exponential backoff with jitter to the client. What changed in the last hour?

## Steps

1. **Symptom (exact):** API responses `429 Too Many Requests`, error body says too many requests. Starts suddenly, often right after a deploy or a new script.

2. **Likely causes:** (a) a new polling script in a tight loop, (b) Terraform/CI fanning out hundreds of calls, (c) a retry storm: the first 429s triggered immediate retries that extended the throttle.

3. What changed in the last hour? New cron, new exporter, new Terraform run. The 429 start time usually matches a deploy.
2. Which endpoint 429s? If it is one endpoint from one client, that client is the offender. If everything 429s, something is hammering broadly.
3. Check for retry-without-backoff in the client code. Immediate retry on 429 is a throttle amplifier.

4. Stop the offender first (kill the loop), then wait out the throttle honoring `Retry-After`.
2. Add exponential backoff with jitter to the client.
3. Cut volume: cache, batch, widen intervals. A dashboard polled every 10s from CI should be a monitor, not a script.

5. **Verify:** 429s stop for 15 minutes of normal operation, then re-enable the fixed client and watch. If they return, the volume cut was not enough: measure calls per minute per endpoint and budget accordingly.

## When to use

You are seeing this: Symptom (exact): API responses 429 Too Many Requests, error body says too many requests. Use this skill when you run into "Diag: 429 rate limited, find the hammering client, backoff, cut volume".

## When not to use

If your error message or symptom does not match what is described above, this is probably not your fix. Search for your exact error text instead of forcing this one to fit.

## Versions

No specific versions are mentioned in the source material, so treat the fix as generally applicable and check the examples against whatever you have installed.

## Why this happens

The original report does not dig into a root cause. It documents the symptom and the fix that resolved it.