Skip to content

python.resilience.missing_graceful_shutdown

Stability High

Detects applications that don’t handle SIGTERM signal for graceful shutdown. Without graceful shutdown, requests are dropped during deployments.

Why It Matters

During deployments, orchestrators like Kubernetes send SIGTERM to allow apps to finish in-flight requests:

Dropped requests — Users see errors during deployments
Data corruption — Half-completed transactions may corrupt data
Resource leaks — Connections aren’t properly closed
Extended downtime — Hard kills after SIGTERM timeout take longer

Example

# ❌ Before (no graceful shutdown)
from fastapi import FastAPI

app = FastAPI()

@app.get("/")
def root():
    return {"message": "Hello"}

# ✅ After (with graceful shutdown using lifespan)
from contextlib import asynccontextmanager
from fastapi import FastAPI
import logging

logger = logging.getLogger(__name__)

@asynccontextmanager
async def lifespan(app: FastAPI):
    # Startup: Initialize resources
    logger.info("Application starting up")
    yield
    # Shutdown: Clean up resources
    logger.info("Shutting down gracefully")
    # Close database connections, flush caches, etc.

app = FastAPI(lifespan=lifespan)

What Unfault Detects

FastAPI/Flask/Django apps without signal handlers
Web applications missing lifespan context managers
Missing atexit registration
Applications without SIGTERM handling

Auto-Fix

For FastAPI applications, Unfault adds a lifespan context manager:

from contextlib import asynccontextmanager

@asynccontextmanager
async def lifespan(app):
    # Startup
    logger.info("Application starting up")
    yield
    # Shutdown
    logger.info("Application shutting down gracefully")

For generic Python applications, it adds signal handlers:

import signal
import sys

def graceful_shutdown(signum, frame):
    """Handle shutdown signals gracefully."""
    logger.info(f"Received signal {signum}, shutting down gracefully...")
    sys.exit(0)

signal.signal(signal.SIGTERM, graceful_shutdown)
signal.signal(signal.SIGINT, graceful_shutdown)

Kubernetes Deployment

# Ensure adequate termination grace period
spec:
  terminationGracePeriodSeconds: 30
  containers:
    - name: app
      lifecycle:
        preStop:
          exec:
            command: ["/bin/sh", "-c", "sleep 5"]  # Allow LB to drain