Thanks for the tips, I’d missed the shared cache thing, and the
read_uncommittedPRAGMA.
I see u are using file-based db - so u don’t need them
More problematic is the most costly operation, which is cleanup
To address the issue with the long cleanup query, you could try optimizing it or splitting it into several smaller queries, and slightly increasing the timeout and pool values.
pool_size: 50,
busy_timeout: 5_000,






















