connection-pooling.md

Connection pooling

from dqliteclient import create_pool

pool = await create_pool(["localhost:9001", "localhost:9002", "localhost:9003"])
async with pool.acquire() as conn:
    rows = await conn.fetch("SELECT 1")

Sizing and startup

The pool opens min_size connections in parallel during initialize(). All initial connects target whichever node the cluster client identifies as the current Raft leader, so the leader serializes its incoming-connection acceptance — raising min_size does not speed up startup linearly with N.

A low default (min_size=1, or a small single digit) keeps cold-start latency predictable. Raise it only when steady-state concurrency genuinely needs warm connections at engine startup.

Where pooling lives in the stack

This pool is for direct dqlite-client use. If you use dqlite-dbapi or sqlalchemy-dqlite, pooling is handled one layer up (SQLAlchemy's QueuePool and its async siblings), and this client-level pool is not used.