bbangert/beaker

Good way or best practices for re-using existing SQLAlchemy engine (connection pool)

benhoyt opened this issue · 3 comments

benhoyt commented

We're using beaker's SQLAlchemy database-backed session support, but our web application also sets up its own SQLAlchemy engine instance. However, we'd like to be able to use a single, shared SQLAlchemy engine instance so that we can have a single SQLAlchemy connection pool, instead of two connection pools (and double the number of PostgreSQL connections as a result).

In ext/database.py it uses the DatabaseNamespaceManager.metadatas dictionary to cache engine instanes, so it looks like we could do something like this to get what we want:

DatabaseNamespaceManager.metadatas[url + table_name] = our_existing_sqla_engine

Is this reasonable? It feels a bit hackish. Is there a better way to use a shared engine? If not, could we add a configuration option for an "engine-fetching function" to allow us to return our existing engine instance?

EndenDragon commented

Any updates on this? I would love to integrate this into my webapp too (my web host limits the number of connections)!

amol- commented

There is currently no "clean" way to pass an existing engine to beaker in ext:database.
The expected way is by relying on ext:sqla instead which actually uses an existing engine.

If anyone is willing to provide a patch to allow it I'll gladly review that, I think we should just get rid of ext:sqla on long term and have a single "sqlalchemy" based extension.