Skip to content

AWS Lambda — Serverless Functions

AWS Lambda is a serverless compute service that lets you run code in response to events without provisioning or managing servers. You pay only for the compute time consumed — there is no charge when code is not running.

In Azure terms: AWS Lambda = Azure Functions

  1. You upload your function code (or write inline)
  2. You configure a trigger (event source)
  3. When the trigger fires, Lambda automatically:
    • Allocates compute resources
    • Runs your function
    • Scales automatically (from 1 to 1,000+ concurrent executions)
    • Reports metrics to CloudWatch
Runtime Versions
Node.js 22.x, 24.x
Python 3.11, 3.12, 3.13, 3.14
Java 8, 11, 17, 21, 25
.NET .NET 10 (dotnet10)
Ruby 3.3, 3.4
Go provided.al2023 (custom runtime)
Custom Runtime Bring your own via Lambda Layers or container images

AWS has announced Node.js 26 and Python 3.15 runtimes for late 2026. Older runtimes are deprecated on a published schedule, so check the Lambda runtimes page before you choose one.

Trigger Type Examples
API API Gateway, Function URL (direct HTTPS)
Storage S3 object created/deleted, DynamoDB Streams
Messaging SQS, SNS, Kinesis, EventBridge
Scheduling EventBridge Scheduler (cron-like)
Auth Cognito User Pools (custom auth flows)
Other CloudFront (Lambda@Edge), ALB, IoT
Setting Description
Memory 128 MB to 10,240 MB (CPU scales proportionally)
Timeout 1 second to 15 minutes (900 seconds)
Concurrency Up to 1,000 concurrent executions per region (soft limit)
Execution Role IAM Role that grants the function permissions
Environment Variables Key-value config passed to function at runtime
Layers Shared dependencies/libraries across functions
VPC Optional — attach to a VPC to access private resources
Metric Price
Requests $0.20 per 1 million requests
Duration $0.0000166667 per GB-second
Free tier 1 million requests/month + 400,000 GB-seconds/month forever

Lambda’s free tier never expires — it’s not limited to the first 12 months.

import json
def lambda_handler(event, context):
# event: the trigger payload (dict)
# context: runtime info (function name, remaining time, etc.)
name = event.get('name', 'World')
return {
'statusCode': 200,
'body': json.dumps(f'Hello, {name}!')
}

A common pattern: API Gateway as the trigger makes Lambda an HTTP endpoint:

Client → API Gateway → Lambda → Response
Terminal window
# Create a Lambda function
aws lambda create-function \
--function-name my-function \
--runtime python3.12 \
--role arn:aws:iam::123456789012:role/lambda-role \
--handler lambda_function.lambda_handler \
--zip-file fileb://function.zip
# Invoke a Lambda function
aws lambda invoke \
--function-name my-function \
--payload '{"name": "AWS"}' \
response.json
# Update function code
aws lambda update-function-code \
--function-name my-function \
--zip-file fileb://function.zip
Scenario Latency Description
Cold start 100ms – 1s+ Lambda spins up a new execution environment
Warm start < 10ms overhead Reuses an existing execution environment

Tips to reduce cold starts:

  • Use Provisioned Concurrency (keeps environments warm, extra cost)
  • Keep deployment packages small
  • Use lighter runtimes (Python, Node.js) for latency-sensitive functions
  • Use Lambda SnapStart (Java only)
Feature AWS Lambda Azure Functions
Max timeout 15 minutes 10 minutes (Consumption), unlimited (Dedicated/Flex)
Max memory 10 GB 1.5 GB (Consumption)
Triggers API GW, S3, SQS, SNS, EventBridge, DynamoDB HTTP, Blob, Queue, Service Bus, Event Hub, Timer
Hosting plans No choice — always serverless Consumption, Premium, Dedicated
Warm instances Provisioned Concurrency Premium (pre-warmed)
Languages Node, Python, Java, .NET, Ruby, Go C#, JS, Python, Java, PowerShell, Go
  • Keep functions small and single-purpose
  • Store secrets in AWS Secrets Manager or SSM Parameter Store, not environment variables
  • Set appropriate timeouts — don’t leave at default 3 seconds for I/O-bound functions
  • Enable X-Ray tracing for distributed debugging
  • Use Dead Letter Queues (DLQ) for failed async invocations
  • Tag functions with environment and team for cost tracking