Please review the logs of mta-hub container.

You'll find query times vary based on dataset, and api call.

Please collect long running queries (greater than 1s)

Run explain on the queries and analyze them, open bugs for relevant poor performing queries.

Please look at the impact of vaccuming the database after large managed import and analysis.

Does query performance improve? Does the overall database size shrink?

This especially relevant after 10k+

review:

https://nomadicsoft.io/ultimate-guide-to-maximizing-sqlite-performance-query-optimization-and-performance-tuning/

https://www.powersync.com/blog/sqlite-optimizations-for-ultra-high-performance

https://phiresky.github.io/blog/2020/sqlite-performance-tuning/