Skip to main content

​
Overview

Database connections are expensive. FluxCRUD uses SQLAlchemy’s connection pooling under the hood but exposes simple configuration options to tune it for high-concurrency workloads.

​
Configuration

You can configure pooling via Flux or Database init: 
flux = Flux(
    app,
    db_url="postgresql+asyncpg://user:pass@localhost/db",
    pool_size=20,       # Base number of connections
    max_overflow=40,    # Max extra connections during spikes
    pool_timeout=30     # Max wait time for a connection
)

​
Best Practices

  • Production: Always use a real pool (not NullPool) for PostgreSQL/MySQL.
  • Serverless: If using AWS Lambda / Azure Functions, consider NullPool if you have an external proxy (like PGBouncer) or keep pool_size small.
  • SQLite: File-based SQLite works best with StaticPool (handled automatically for in-memory) or strictly serialized access. FluxCRUD handles safe defaults.