fix(db): honor SQLite pool settings

[onatozmenn]

HUMAN:

I tested both SQLite memory URL formats locally and ran the test suite to make sure nothing broke.

AGENT:

File-backed SQLite was silently using SQLAlchemy's 5+10 pool defaults, so changing the service's pool settings had no effect. This passes the configured size, overflow, and timeout through while leaving in-memory databases on StaticPool.

I covered the regular :memory: URL and both SQLite URI memory forms so the new kwargs don't break test or embedded setups.

Tests: uv run pytest tests/test_db.py -q (31 passed), Ruff format/check, and Pyright. The wider non-integration suite reached 1,439 passes; its only Postgres testcontainer setup timeout passed on an isolated rerun.

Fixes #366

[all-hands-bot]

👋 This PR needs a couple of things fixed before OpenHands can review it:

• the PR description's HUMAN: section needs at least 20 characters describing what you tested, not just the template placeholder

Push an update once this is addressed and this check re-runs automatically.

This is an automated check - no AI was used to generate this comment.