villebro commented on issue #30064: URL: https://github.com/apache/superset/issues/30064#issuecomment-2333175111
> > I think it would be a better idea to provide sane defaults in the extensions of BaseEngineSpec > > > > Mmmh, I get the idea, but I'd say the `db_engine_spec` abstraction is there to serve analytics-type workloads. I'd say for analytics workload on postgres you'd want AUTOCOMMIT, as typically you want as little as possible isolation overhead. Very different from OLTP. Say if you happen to use Postgres for both metadata db and analytics workload, you'd probably want `READ COMMITTED` for the metadata db and AUTOCOMMIT for analytics... > > > > I'd open PR with a dynamic default for `config.py`, only specifying `READ COMMITTED` for `mysql` and `postgres`, but realizing that at `config.py` parsing-time we may not know what the database connection look like. Maybe using default logic in `DB_CONNECTION_MUTATOR` (?). Or should we just not specify `SQLALCHEMY_ENGINE_OPTIONS` in `config.py`? Yeah you're absolutely right, my bad 👍 However, I'm still not entirely buying into having this in `config.py`, but let me sleep on it and see if I'm wiser in the morning. -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected] --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
