Bring your towel and join one of the fastest growing data communities. Welcome to our second-generation open source orchestration platform, a completely rethought approach to dataflow automation.

Prefect Community

Hi - does anyone know where i can the number of API requests i make towards the cloud server?

Last time I asked it wasn't available and I haven't been able to find any information about what even counts towards the limit

It's becoming more of a requirement for us that we can figure this out so that we know when we're risking killing things in production :shrug:

Okay, thank you :slightly_smiling_face: . I agree, it would be great to see whether we are close to breaching the API limit.

hey <@U0675Q5B402> -- attaching our rate limit <https://docs.prefect.io/latest/cloud/rate-limits/#:~:text=400%20per%20minute%20for%20personal,per%20minute%20for%20Pro%20accounts|documentation>!

<@U03VD2Q7DP1>, along with any API calls from within your code, the creation of a task and or flow will count towards the flow, flow run, and task rate limits.

<@U062C5BHJEA> Is there any documentation about what parts of the code make what types of API calls? We're using the prefect client rather than making API calls directly so there's a layer of abstraction that is obviously useful for our code but makes any effort on our side to figure out where specific API calls are happening much more involved

I've also looked at the concurrency limits for the work pool and queues and everything reads that this is only set at the flow run level, so this has limited use if you have flows with many tasks

I've looked at the tag based concurrency limit for tasks, but that stops more tasks moving into the `Running` state and still creates all the tasks at the same time so as far as I can tell doesn't actually stop us hitting the rate limit because all the creation events happen at the start (it might help us not hit the rate limit multiple times due to changes of state, but the actual creation events are still happening)

Are there any other avenues I can try that would reduce simultaneous API hits (and therefore the rate limit)?

<@U03VD2Q7DP1> Other than concurrency limits, you can attempt to adjust the environment variable that increases the <https://docs.prefect.io/2.14.5/api-ref/prefect/settings/#prefect.settings.PREFECT_CLIENT_MAX_RETRIES|max number of retries with a 429>.

Let me gather some more info and follow up with a final response!

Ok so final confirmation that I've checked <@U0675Q5B402> <@U03VD2Q7DP1> -- Rate limits are hit through the <https://app.prefect.cloud/api/docs?deviceId=0f152d6a-01c5-4493-bc1e-0c2918410fa8#tag/Task-Runs/operation/create_task_run_api_accounts__account_id__workspaces__workspace_id__task_runs__post|Create Task Run> and <https://app.prefect.cloud/api/docs?deviceId=0f152d6a-01c5-4493-bc1e-0c2918410fa8#tag/Flow-Runs/operation/create_flow_run_api_accounts__account_id__workspaces__workspace_id__flow_runs__post|Create Flow Run> calls. You can see these in your logs as well. So, for example, if you have something that maps over a task / for loop / etc. each time it creates a task run it will count towards the rate limit.

If you are hitting these and have attempted concurrency and retries, the best option is to then increase your rate limit, which comes with an account upgrade.

<@U062C5BHJEA> - okay - thank you for clarifying :slightly_smiling_face: