AWS Lambda Python Integration

Patterns for creating high-performance AWS Lambda functions in Python with optimized cold starts and clean architecture.

Overview

AWS Lambda Python integration with two approaches: AWS Chalice (full-featured framework) and Raw Python (minimal overhead). Both support API Gateway/ALB integration with production-ready configurations.

When to Use

Use this skill when:

• Creating new Lambda functions in Python
• Migrating existing Python applications to Lambda
• Optimizing cold start performance for Python Lambda
• Choosing between framework-based and minimal Python approaches
• Configuring API Gateway or ALB integration
• Setting up deployment pipelines for Python Lambda

Instructions

1. Choose Your Approach

Approach	Cold Start	Best For	Complexity
AWS Chalice	< 200ms	REST APIs, rapid development, built-in routing	Low
Raw Python	< 100ms	Simple handlers, maximum control, minimal dependencies	Low

2. Project Structure

AWS Chalice Structure

my-chalice-app/
├── app.py                    # Main application with routes
├── requirements.txt          # Dependencies
├── .chalice/
│   ├── config.json          # Chalice configuration
│   └── deploy/              # Deployment artifacts
├── chalicelib/              # Additional modules
│   ├── __init__.py
│   └── services.py
└── tests/
    └── test_app.py

Raw Python Structure

my-lambda-function/
├── lambda_function.py       # Handler entry point
├── requirements.txt         # Dependencies
├── template.yaml            # SAM/CloudFormation template
└── src/                     # Additional modules
    ├── __init__.py
    ├── handlers.py
    └── utils.py

3. Implementation Examples

See the References section for detailed implementation guides. Quick examples:

AWS Chalice:

from chalice import Chalice
app = Chalice(app_name='my-api')

@app.route('/')
def index():
    return {'message': 'Hello from Chalice!'}

Raw Python:

def lambda_handler(event, context):
    return {
        'statusCode': 200,
        'body': json.dumps({'message': 'Hello from Lambda!'})
    }

Core Concepts

Cold Start Optimization

Key strategies:

1. Initialize at module level - Persists across warm invocations
2. Use lazy loading - Defer heavy imports until needed
3. Cache boto3 clients - Reuse connections between invocations

See Raw Python Lambda for detailed patterns.

Connection Management

Create clients at module level and reuse:

_dynamodb = None

def get_table():
    global _dynamodb
    if _dynamodb is None:
        _dynamodb = boto3.resource('dynamodb').Table('my-table')
    return _dynamodb

Environment Configuration

class Config:
    TABLE_NAME = os.environ.get('TABLE_NAME')
    DEBUG = os.environ.get('DEBUG', 'false').lower() == 'true'

    @classmethod
    def validate(cls):
        if not cls.TABLE_NAME:
            raise ValueError("TABLE_NAME required")

Best Practices

Memory and Timeout Configuration

• Memory: Start with 256MB for simple handlers, 512MB for complex operations
• Timeout: Set based on expected processing time
  • Simple handlers: 3-5 seconds
  • API with DB calls: 10-15 seconds
  • Data processing: 30-60 seconds

Dependencies

Keep requirements.txt minimal:

# Core AWS SDK - always needed
boto3>=1.35.0

# Only add what you need
requests>=2.32.0  # If calling external APIs
pydantic>=2.5.0   # If using data validation

Error Handling

Return proper HTTP codes with request ID:

def lambda_handler(event, context):
    try:
        result = process_event(event)
        return {'statusCode': 200, 'body': json.dumps(result)}
    except ValueError as e:
        return {'statusCode': 400, 'body': json.dumps({'error': str(e)})}
    except Exception as e:
        print(f"Error: {str(e)}")  # Log to CloudWatch
        return {'statusCode': 500, 'body': json.dumps({'error': 'Internal error'})}

See Raw Python Lambda for structured error patterns.

Logging

Use structured logging for CloudWatch Insights:

import logging, json
logger = logging.getLogger()
logger.setLevel(logging.INFO)

# Structured log
logger.info(json.dumps({
    'eventType': 'REQUEST',
    'requestId': context.aws_request_id,
    'path': event.get('path')
}))

See Raw Python Lambda for advanced patterns.

Deployment Options

Quick Start

Validation Checkpoint: Always run serverless print or sam validate before deploying to catch configuration errors early.

Serverless Framework:

# serverless.yml
service: my-python-api
provider:
  name: aws
  runtime: python3.12  # or python3.11
functions:
  api:
    handler: lambda_function.lambda_handler
    events:
      - http:
          path: /{proxy+}
          method: ANY

AWS SAM:

# template.yaml
AWSTemplateFormatVersion: '2010-09-09'
Transform: AWS::Serverless-2016-10-31

Resources:
  ApiFunction:
    Type: AWS::Serverless::Function
    Properties:
      CodeUri: ./
      Handler: lambda_function.lambda_handler
      Runtime: python3.12  # or python3.11
      Events:
        ApiEvent:
          Type: Api
          Properties:
            Path: /{proxy+}
            Method: ANY

AWS Chalice:

chalice new-project my-api
cd my-api
chalice local 8080  # Test locally before deploying
chalice deploy --stage dev

Validation Checkpoint: Test locally with chalice local or sam local invoke before deploying to production.

For complete deployment configurations including CI/CD, environment-specific settings, and advanced SAM/Serverless patterns, see Serverless Deployment.

Constraints and Warnings

Lambda Limits

• Deployment package: 250MB unzipped maximum (50MB zipped)
• Memory: 128MB to 10GB
• Timeout: 15 minutes maximum
• Concurrent executions: 1000 default (adjustable)
• Environment variables: 4KB total size

Python-Specific Considerations

• Cold start: Python has excellent cold start performance; avoid heavy imports at module level
• Dependencies: Keep requirements.txt minimal; use Lambda Layers for shared dependencies
• Native dependencies: Must be compiled for Amazon Linux 2 (x86_64 or arm64)

Common Pitfalls

1. Importing heavy libraries at module level - Defer to function level if not always needed
2. Not handling Lambda context - Use context.get_remaining_time_in_millis() for timeout awareness
3. Not validating input - Always validate and sanitize event data
4. Printing sensitive data - Be careful with logs and CloudWatch

Error Recovery: If deployment fails, check