A lot of the times, when I try to open a task-run, the page loads for 5 minutes then redirects me to

getting-started

. For our particular usage, this makes Prefect unusable, as we need to be able to inspect processes quickly (or even at all). Is there a common issue that causes for this?

Can you provide more details? 1. Is it on Prefect Cloud or Server? 2. Does it happen only with a specific flow or all flows? 3. Does it happen only in a specific browser or all browsers show the same latency? 4. Can you run “prefect diagnostics” and share the output?

Prefect Server, with Helm in particular. Happens with every flow and every browser. How do I run prefect diagnostics?

You would need to exec to the pod to run it, but it’s not that relevant. How are the components deployed? Is everything within the same network to minimize latency? In general, with Prefect Cloud we provide some performance optimizations because we control the underlying infrastructure. For Server, you would need to dive deeper to investigate what might be the bottleneck in your infrastructure. Perhaps you could leverage Prometheus to monitor the pods for all components and see what might be happening this way?