Oliver Klein

Same issue, thanks for posting a workaround! Hope to see this fixed

I would also support the request to support arrow datatypes which I guess become the new normal in Pandas 2. My current workaround is to convert the arrow dtypes to...

Confirm that problem is gone with #1125

Please note that the use/support of pyproject.toml is not bound to poetry, as pyproject.toml is adopted outside of poetry these days. https://packaging.python.org/en/latest/guides/writing-pyproject-toml/

I have the same problem currently, but with 2.20.15. I realized it when moving from 2.20.15-python3.10 to 2.20.15-python3.11. However, it stays to be a problem even when I now rollback....

> I am seeing exactly this behavior -- timeout errors with messages about things taking longer than their loop allowances, and then Prefect's workflow management becoming entirely inactive -- but...

> > I am seeing exactly this behavior -- timeout errors with messages about things taking longer than their loop allowances, and then Prefect's workflow management becoming entirely inactive --...

One observation when running `SELECT datname, state, count(*) FROM pg_stat_activity group by datname, state order by count desc;` the clear winner is the prefect database with in my case currently...

``` datid | datname | pid | leader_pid | usesysid | usename | application_name | client_addr | client_hostname | client_port | backend_start | xact_start | query_start | state_change | wait_event_type...

Another finding I could add: we have 2 instances in cloud DEV and PROD. On the seldomly used DEV environment I see 5 such idle connections, so much less that...