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

Did active try to use Yugabyte instead of postgres with Prefect Server 1.x?
Looking to get more resiliency and hot hot availability.

The easiest would be to give it a try. I'm always a bit skeptical when I see "Postgres-compatible" - it's still not Postgres?

Where are you running this? I can totally understand the problem of ensuring resiliency, but have you looked at the Postgres managed services from cloud providers? they seem to provide features for more resiliency, not sure about hot-hot availability (multi-AZ RDS?)

I would love to have access to cloud providers and DBaaS, however, the journey is only starting and the movement is slow.
At the moment we have everything on Kubernetes across multiple cluster, however postgres is not simple to setup for full high availability.

I will give it a try and let you know how it goes. I never tried it, however as this is certified as Postgres compatible by Hasura 2.x I want to give it a try.
I will let you know the feedback.
Do you know if Prefect server queries directly the Db or everything is handled using Hasura

I understand, migrating from on-prem to :cloud: is not easy 

Good question, let me ask the team

btw do you plan to actually move to Prefect 2.0? :wink:

with Prefect 2.0, I checked with Bill and it works based on ORM and SQLAlchemy, so doing this should work in theory (I haven't tested that):
```export PREFECT_ORION_DATABASE_CONNECTION_URL="<postgresql+asyncpg://yb_db_user:yb_pwd@yb_host:5433/db_name>"```
this would work because YougabyteDB <https://docs.yugabyte.com/preview/integrations/sqlalchemy/|supports SQLAlchemy>

Yes I'm planning to migrate to prefect 2 however I need to be out of the beta stage. Looking forward about that

<@U012J275170> I asked the team about YugabyteDB with Prefect Server and here is the answer:
• in 1.0, almost everything goes through Hasura, so there is some chance that this might work and it would be best if you give it a try and report back - I'd appreciate that!
• there are some places where we query the DB directly using `asyncpg`, and we also run migrations via Alembic, so those use cases don't go through Hasura and you would need to be more careful in those use cases
So the bottom line is: there is no dedicated support for YugabyteDB but with some work on your end it _could_ work, but no guarantee.

But from a long-term perspective, you may be better off if you start using it with Prefect 2.0 - due to the ORM layer rather than Hasura, this is much easier