Skip to content

python.db.missing_timeout

Stability High

Detects database connections and queries that don’t have proper timeout configuration, which can lead to hanging connections and cascading failures.

Why It Matters

Database connections without timeouts can:

Hang indefinitely — Network issues can cause connections to wait forever
Exhaust connection pools — Stuck connections are never returned to the pool
Cause cascading failures — Application workers get blocked waiting for DB
Prevent graceful shutdown — Long-running queries ignore shutdown signals

Example

# ❌ Before (no timeout)
from sqlalchemy import create_engine

engine = create_engine("postgresql://localhost/db")

# ✅ After (with timeout)
from sqlalchemy import create_engine

engine = create_engine(
    "postgresql://localhost/db",
    connect_args={"connect_timeout": 10},
    pool_pre_ping=True,  # Also recommended
)

What Unfault Detects

SQLAlchemy create_engine() without connect_args timeout
SQLAlchemy create_async_engine() without timeout
psycopg2/psycopg connect() without connect_timeout
asyncpg connect() without timeout parameter

Auto-Fix

Unfault generates patches that add appropriate timeout parameters:

# SQLAlchemy (sync)
engine = create_engine(
    DATABASE_URL,
    connect_args={"connect_timeout": 10},
)

# SQLAlchemy (async with asyncpg)
engine = create_async_engine(
    DATABASE_URL,
    connect_args={"timeout": 10},  # asyncpg uses 'timeout'
)

# psycopg2
conn = psycopg2.connect(dsn, connect_timeout=10)

# asyncpg
conn = await asyncpg.connect(dsn, timeout=10)

Best Practices

# Comprehensive SQLAlchemy configuration
engine = create_engine(
    DATABASE_URL,
    connect_args={
        "connect_timeout": 10,
        "options": "-c statement_timeout=30000",  # 30s query timeout
    },
    pool_pre_ping=True,  # Verify connections before use
    pool_recycle=3600,   # Recycle connections hourly
    pool_timeout=30,     # Max wait for connection from pool
)