- We just started getting 429 Client error '429 To...
# ask-community
a
• We just started getting 429 Client error '429 Too Many Requests' inside our prefect cloud instance. Is there a way to clearly see the calls per minute. Our tier is at
2,000 per minute for Pro accounts
so I doubt we are hitting it unless we have hidden hanging flows somewhere
👀 1
b
Hey Andy! Here's a bit of a copy/pasta rate limiting:
For the rate limits, any call the is made to the flow_run, task_run, and flow endpoints counts towards the limit. This includes task/flow run state changes, task/flow run creation, deletion, cancellation, etc.
It bears mentioning that the 429 errors are automatically retried by the client a certain number of times. Retries can also be set to a configurable number (see here ), which can help with managing smaller spikes in requests. However, consistent high volumes of requests would consume a lot of bandwidth, and retries wouldn't be of much help on that front. At that point, you may consider spreading out the number of flow runs and task runs that run at a given time.
You can add sleeps to your tasks, or set concurrency limits to avoid bumping into the rate limit.
Work pool concurrency
Task run concurrency
I believe that there will be something in the works for summarzing API requests, and surfacing that in the UI. A user created a feature request here that you can follow along with
With that being said, we'd be happy to investigate this on our end. Would you mind sharing the account slug with us, and whether or not you're using concurrency tags on tasks?
a
ill dm