Terraform Module tag-based-lambda-warmer

A Terraform module for deploying an automated Lambda function warming solution. This module helps prevent cold starts in your Lambda functions by periodically invoking them based on tags.

Overview

This module creates a Lambda function that automatically warms up other Lambda functions based on specified tags. It uses EventBridge Scheduler to trigger the warming process at regular intervals.

Features

• Tag-based function selection for warming
• Configurable warming schedule
• Customizable tag key/value pairs for targeting functions
• Customizable Lambda Warmer function invocation type ( Event / RequestResponse )

Usage

module "lambda_warmer" {
  source = "../modules/lambda_warmer"

  aws_region = "ap-northeast-1"

  # Optional: Custom configuration
  environment = "production"
  prewarm_tag_key = "Project"
  prewarm_tag_value = "MyProject"
  lambda_schedule_expression = "rate(5 minutes)"
  timeout = 60
  memory_size = 256
  scheduler_max_retry_attempts = 0
  invocation_type = "Event"
}

Tagging Lambda Functions for Warming

To mark a Lambda function for warming, add the appropriate tags:

resource "aws_lambda_function" "example" {
  # ... other configuration ...

  tags = {
    Prewarm = "true"  # Default tags
    # Or use custom tags as configured in the module
    Project = "MyProject"
  }
}

Code Snippet for Handling Prewarming in Your Function

When a Lambda function is invoked for prewarming, it's important to differentiate prewarming requests from normal business logic. Below is an python example of how to handle prewarming requests

def lambda_handler(event, context):
    # Identify if the incoming event is a prewarm request
    if event.get("action") == "PREWARM":
        logger.info("Received a prewarm request. Skipping business logic.")
        return {
            "statusCode": 200,
            "body": "Successfully warmed up"
        }

Requirements

Name	Version
Terraform	>= 1.8.0
AWS Provider	>= 5.54

Variables

Required Variables

Name	Description	Type
aws_region	AWS region where the Lambda warmer will be deployed	string

Optional Variables

Name	Description	Type	Default
environment	Current environment, e.g. dev, stage, prod	string	"default"
lambda_function_name	Name of the Lambda function used for prewarming	string	"lambda-warmer"
lambda_schedule_expression	EventBridge schedule expression for triggering the Lambda warmer	string	"rate(5 minutes)"
scheduler_max_retry_attempts	Maximum retry attempts for the EventBridge scheduler target	number	0
timeout	Timeout for the Lambda function in seconds	number	10
memory_size	Memory size for the Lambda function	number	128
prewarm_tag_key	The tag key to identify functions that need prewarming	string	"Prewarm"
prewarm_tag_value	The expected value of the tag for functions that need prewarming	string	"true"
invocation_type	The invocation type of the Lambda Warmer function (Event / RequestResponse)	string	"Event"

Outputs

Name	Description
scheduler_group_arn	ARN of the EventBridge Scheduler Group
scheduler_arn	ARN of the EventBridge Scheduler
lambda_function_name	Name of the Lambda warmer function
lambda_function_arn	ARN of the Lambda warmer function
lambda_role_name	Name of the Lambda IAM Role
lambda_role_arn	ARN of the Lambda IAM Role

Directory Structure

.
├── README.md
├── main.tf                # Main Terraform configuration
├── variables.tf           # Input variables
├── versions.tf            # Terraform version constraints
├── outputs.tf             # Output definitions
└── lambda_warmer/         # Lambda function code
    └── lambda_function.py # Python implementation

IAM Roles and Permissions

The module creates two separate IAM roles:

1. Lambda Role:

   • CloudWatch Logs access
   • Lambda function invocation

2. EventBridge Scheduler Role:

   • Permission to invoke the warmer Lambda function

Schedule Expression Examples

• Every 5 minutes: rate(5 minutes)
• Every hour: rate(1 hour)

Monitoring and Logs

The Lambda warmer function logs its activities to CloudWatch Logs. You can monitor:

• Functions being warmed
• Warming success/failure
• Number of functions processed

Contact

If you encounter any issues, feel free to reach out at [email protected] or submit an issue.

License

This project is licensed under the MIT License. See the LICENSE file for details.

Copyright © 2024 AWS Educate Cloud Ambassador Taiwan