Replies: 2 comments 1 reply
-
|
Working on it - will publish a newer version soon. |
Beta Was this translation helpful? Give feedback.
0 replies
-
|
Should be fixed on 0.24.0 |
Beta Was this translation helpful? Give feedback.
1 reply
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Uh oh!
There was an error while loading. Please reload this page.
-
Something about our upgrade from 0.21 -> 0.23 caused significant performance degradation in our app:
Has anyone experienced anything similar and how did they deal with it?
We saw that the 0.21->0.22 upgrade migrated fully to SQLAlchemy 2.0 and dropped the
databaseslibrary. It looks like the default behavior now wraps standalone queries in an implicit BEGIN/COMMIT, and we strongly suspect this is giving us excess commit overhead.We thought we could set our SQL engine's
isolation_levelto "AUTOCOMMIT" to put us back to the previous legacy behavior. We're now seeing a bunch of these errors when ormar tries to use savepoints internally:sqlalchemy.dialects.postgresql.asyncpg.AsyncAdapt_asyncpg_dbapi.Error: <class 'asyncpg.exceptions.NoActiveSQLTransactionError'>: SAVEPOINT can only be used in transaction blocksHas anyone else successfully patched in
AUTOCOMMITto restore performance, or is the officially recommended path just biting the bullet and wrapping everything in explicitasync with database.transaction()? Or how else did people deal with this if encountered? Any advice is appreciated!Beta Was this translation helpful? Give feedback.
All reactions