# rfp agent hit rate limit mid compliance check and failed to finish: how to fix it

An RFP agent hitting a rate limit mid compliance check needs checkpointed checks and paced retries. Save each completed check, back off on the limit, and resume. Compliance checks are numerous and bursty, which is exactly the pattern limits punish.

```text
rfp agent hit rate limit mid compliance check and failed to finish
```

## When to use this skill
- Compliance check runs die on rate limits
- Restarts recheck already-checked items

## When not to use this skill
- Checks fail on content rather than limits (a check logic issue)
- The limit is permanent (a quota conversation)

## Tool compatibility
Agent compliance checking against rate-limited model APIs.
Related area: model APIs and inference quotas.
Cloud UIs change labels over time; if a button moved, search settings for the closest match.

## Before you start
- Sign in to the tool and keep one tab open so sessions stay clean
- Copy the exact error text and note the time it happened
- Set aside 15 to 20 minutes; most of these resolve in a single pass

## Quick diagnosis
Make sure this is the right skill before you invest the 15 minutes.
- If you are dealing with compliance check runs die on rate limits, start at step 1 below.
- If instead checks fail on content rather than limits (a check logic issue), stop here; you need a different skill.
- If neither quite matches, compare your screen against the verbatim error block at the top; that block is exactly what this skill covers.

## Fix it in steps
### 1. Checkpoint each check

```
Save the result of every compliance check as it completes.
```
Expected: a resume skips finished checks.

### 2. Add backoff on limit errors

```
Pause and retry with exponential backoff when the limit hits.
```
Expected: the run survives limit windows.

### 3. Pace the check calls

```
Space checks out to stay under the quota.
```
Expected: fewer limit hits overall.

### 4. Resume to a full pass

```
Rerun and let it complete the checklist.
```
Expected: every requirement checked.

## Confirm each step worked
Run through this checklist before moving on. If a check fails, redo that step instead of pushing ahead.
- Step 1 (checkpoint each check): Expected: a resume skips finished checks.
- Step 2 (add backoff on limit errors): Expected: the run survives limit windows.
- Step 3 (pace the check calls): Expected: fewer limit hits overall.
- Step 4 (resume to a full pass): Expected: every requirement checked.

## Quick recap
1. Checkpoint each check
2. Add backoff on limit errors
3. Pace the check calls
4. Resume to a full pass

## Still stuck?
Gather three things: the exact error text, the time it happened, and what changed just before it broke.
Check the vendor status page for an active incident. Then open a support ticket with those three things;
it cuts the back-and-forth in half. For agent runs, attach the last 50 lines of the agent log too,
since the loop or timeout signature in the log usually names the cause.

## Variant phrasings
### rate limit mid compliance check
Checkpoint plus pacing. If this matches what you saw, the steps above apply as written.

### agent failed to finish compliance
Bursty checks need pacing by design. If this matches what you saw, the steps above apply as written.

## Why it happens
Compliance runs fire dozens of model calls in quick succession. The limit error kills the run statelessly, so without checkpoints every retry replays the burst and hits the wall again.

## Edge cases
- Group related checks into fewer calls where possible
- Track per-minute vs per-day quotas separately
- Log check progress so partial runs are auditable

## Provenance

Resolved from the public thread: https://vectle.com/posts/pst_JRSJfFXySpB959JK1NbWtA