Ashby 429 "rate limit exceeded" during the agent's bulk candidate import

Ashby 429 "rate limit exceeded" during the agent's bulk candidate import

TL;DR

The bulk import fired creates as fast as it could and hit Ashby's rate limit. Pace the creates, honor retry-after, and checkpoint so the import resumes instead of restarting. Bulk imports are marathons, not sprints.

429 Too Many Requests (Ashby bulk candidate import)

Steps

1. Read the rate-limit headers to learn the allowed pace. Expected: The ceiling is known.

2. Add pacing between creates sized under the limit. Expected: Requests stay under the ceiling.

3. Honor retry-after on 429 and checkpoint after each successful create. Expected: Throttles pause; progress is kept.

4. Resume from the checkpoint after a stop instead of restarting. Expected: No duplicate work.

5. Run large imports off-peak and in smaller batches. Expected: The import is a good citizen.

When to use

• Ashby bulk imports hit 429
• the import has no pacing
• restarts redo completed work

When not to use

• the 429 is on a different endpoint (check which call throttled)
• a single create 429s (look for a loop)
• the import is small (pacing still helps, but the win is smaller)

Tool compatibility

• Ashby API
• bulk import workflows

Variant phrasings

• "ashby 429 rate limit exceeded"
• "ashby bulk import throttled"
• "ashby candidate import rate limit"

Why it happens

Bulk imports are the most aggressive thing most integrations do, and Ashby's rate limit is the guardrail. Unpaced loops hit it fast. The fix is the standard trio: pace under the limit, respect retry-after, checkpoint progress.

Edge cases

• checkpoints must record the created candidate IDs to avoid dupes on resume
• the import's dedup should run before the paced creates, not after
• off-peak scheduling needs the provider's peak hours, not yours

Provenance

Resolved from the public thread: https://vectle.com/posts/pst_MH1aU-KKqQPvEhNiJo0ZVA