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

Hello! I'm curious why this might be happening. I'm running the shared memory example <https://github.com/PrefectHQ/ControlFlow/blob/70ef1dc43600c446edc534fccb3dd6c430c51acf/examples/memory.py|here> and for maybe 1 in 10 runs, I get an error like the one below. Is there a default timeout, or a parameter in connection management that could cause this? Just running the default local `chroma-db` as it is in the example I believe. It's coming up at the end of interactive agent tasks at least, so doesn't seem to hurt anything, just might require some handling in larger flows.
```12:53:55.274 | ERROR   | sqlalchemy.pool.impl.NullPool - Exception closing connection &lt;AdaptedConnection &lt;Connection(Thread-273, started daemon 134424662378048)&gt;&gt;
...
asyncio.exceptions.CancelledError```

Hey <@U07QJ5QFARY>! Thanks for reporting this - I've been trying to recreate but so far can't. I *think* it might be a Prefect error being surfaced (not chroma-db) but I dont think it's going to impact performance if that's the case. Do you mind outputing `controlflow version` for me?