Hi! Question about caching - here all tasks have finished, but only one has "duration". Does it mean that all the task that finished, are in one of the successful state, but don't have "duration" were read from cache?

Hey @Krzysztof Nawara, will ask the team tom (and also the other question)

How do I run flow with debug-level logs? I set export PREFECT__LOGGING__LEVEL=DEBUG for the terminal from which I'm running

prefect server start

but it doesn't seem to be enough. Do I need to specify it for agent/flow when submitting/ both?

The agent logs and flow logs are separate. I think it’ll be easier to do it from the run config like

LocalRun(env={"PREFECT__LOGGING__LEVEL": "DEBUG"})

for the Flow side, and for the agent it would be

prefect agent local start --log-level=DEBUG

. It’s a bit harder if you are using the Python interface to spin up the agent. You need to modify the

config.toml

with:

[cloud.agent]
level = "debug"

Okay, so to be able to see DEBUG level logs in Prefect UI it's enough to set _export PREFECT__LOGGING__LEVEL=DEBUG_ for my Python script which will be submitting (calling flow.submit()) flow? And then that version of a flow will always be executed with DEBUG logging?

A bit unclear what you’re asking. The environment variable

PREFECT___LOGGING___LEVEL

needs to be on the agent that is executing the Flow. If you export it and register on a different machine, that environment variable will not be registered along with your Flow. The way to attach it to you Flow through be through the RunConfig like

LocalRun(env={"PREFECT__LOGGING__LEVEL": "DEBUG"})

, which would then set the environment variable when you run your flow.

Okay, I tried setting the env variable for the flow, let me try with the RunConfig

I think the

log-level

flag is not available for all types of agents.