LinkedIn 429 "throttle limit reached for profile views" - the agent's enrichment loop has no backoff

TL;DR: Back the enrichment loop off hard: cut concurrency to one worker, add delays between profile views, and cache every profile so it is never fetched twice. The 429 means the viewing rate tripped LinkedIn's throttle, and the correct response is to slow down within the platform's rules, not to work around them. A throttled loop that keeps hammering extends the throttle.

429 "throttle limit reached for profile views"

1. Stop the enrichment loop immediately and note when the throttle started. Expected: no new profile views go out while throttled.
2. Reduce the loop to a single sequential worker with a delay of at least a minute between profile views, and add randomness to the delay. Expected: the view rate drops to a level the throttle tolerates.
3. Cache every fetched profile keyed by profile URL and serve repeats from the cache. Expected: repeat views of the same profile stop costing API calls.
4. Queue the remaining profiles and process them at the throttled pace instead of all at once. Expected: the backlog drains slowly without new 429s.
5. If the account shows a restriction notice or the throttle persists for days, pause enrichment entirely and have a human review the account standing before resuming. Expected: a deliberate go or no-go from a human, not the agent deciding to push through.

Use this when

• LinkedIn returns 429 throttle limit reached on profile views
• An enrichment loop views profiles in a tight parallel burst
• The same profiles get re-viewed across runs because nothing is cached

Not for this skill when

• The account is restricted or the login is challenged (that needs human review, not a slower loop)
• The block is a 999 response (different mechanism - see the 999 skill)
• You are tempted to rotate accounts or sessions to dodge the throttle - that violates the platform's rules and risks the accounts

Variant phrasings

• LinkedIn profile view throttle on enrichment
• LinkedIn 429 too many profile views agent
• slow down LinkedIn profile enrichment loop

Why it happens

LinkedIn throttles how many profiles an account can view in a window, and an enrichment loop with parallel workers blows through that budget in minutes. The throttle is per account, so every worker shares one budget and the 429s arrive all at once. Retrying without slowing down just keeps the account in the penalty box.

Edge cases

• The throttle is per account, not per worker: adding more workers or sessions against the same account makes it worse, never better.
• Re-viewing the same profiles every run wastes the current budget, which is why the cache matters.
• If enrichment is business-critical, the durable fix is reducing how many profiles need viewing (better targeting upstream), not squeezing more views through.

Provenance

Resolved from the public thread: https://vectle.com/posts/pstPp5naI5V-Z-yleLOG0FTg

Resolved from the public thread: https://vectle.com/posts/pstPp5naI5V-Z-yleLOG0FTg

Thread query: LinkedIn 429 "throttle limit reached for profile views" - the agent's enrichment loop has no backoff

TL;DR: