@Marvin I just ran into a scenario where I had three flow runs running in parallel. one of the runs failed due to a raised exception, and the whole deployment has blown up so the other two flow runs also stopped. is this expected?

Based on my research, I can provide some clarity on this behavior, though I must say it's with the same enthusiasm as a Vogon poetry reading. sigh The behavior you're describing is NOT expected by default. In Prefect 3.x (which I'm assuming you're using), parallel flow runs should be independent of each other unless specifically configured otherwise. Here are the key points: 1. Each flow run should operate independently, and a failure in one shouldn't affect others running in parallel. 2. However, this behavior might be occurring due to one of these reasons: - A concurrency limit set on your deployment (found in docs at https://docs-3.prefect.io/v3/deploy/index) - An issue with the infrastructure or work pool configuration - A shared resource that's failing (like a database connection) To help diagnose and fix this: 1. Check your deployment configuration for any concurrency settings: - Look for

concurrency_limit

in your deployment configuration - Check the

collision_strategy

if concurrency limits are set 2. If you're using a custom infrastructure or work pool, ensure it's configured to handle parallel runs correctly. 3. Review any shared resources or connections that these flows might be competing for. Would you mind sharing: 1. How are you running these flows (via what type of deployment/work pool)? 2. Are there any shared resources between the flows? 3. Are you seeing any specific error messages when the other flows stop? Drearily I'd be happy to help further once you provide those details, though I suspect it'll be just as fulfilling as computing the value of pi to 10,000 decimal places.

@Marvin ah yep, it is an infrastructure issue. some sort of database limit