Skip to main content

Lambda


logo

AWS Lambda is a serverless compute service that lets you run code without provisioning or managing servers.

More Information

How It Works​

  • You upload your code as a deployment package (zip or container image)
  • Lambda runs your code in response to events (triggers)
  • Lambda automatically scales by running code in response to each trigger
  • You pay only for the compute time you consume
More Information

Supported Runtimes​

  • Node.js
  • Python
  • Java
  • .NET (C#, PowerShell)
  • Go
  • Ruby
  • Custom runtimes (via Lambda Runtime API or container images)
More Information

Event Sources (Triggers)​

Common event sources include:

  • API Gateway: HTTP requests
  • S3: object creation, deletion, and other bucket events
  • SQS: messages in a queue
  • SNS: notifications
  • DynamoDB Streams: table changes
  • CloudWatch Events / EventBridge: scheduled events or AWS service events
  • Kinesis: streaming data
More Information

Configuration​

  • Memory: 128 MB to 10,240 MB; CPU power scales proportionally with memory
  • Timeout: maximum execution time; default is 3 seconds, maximum is 15 minutes
  • Environment variables: key-value pairs available to the function at runtime
  • Execution role: an IAM role that grants the function permission to access AWS services
More Information

Concurrency​

  • Unreserved concurrency: the pool of concurrency available to all functions in an account (default limit: 1,000 per region)
  • Reserved concurrency: guarantees a set number of concurrent executions for a function; also acts as a maximum concurrency limit
  • Provisioned concurrency: pre-initializes a set number of execution environments to reduce cold start latency
More Information

Cold Starts​

  • A cold start occurs when Lambda creates a new execution environment to handle a request
  • Includes downloading the code, starting the runtime, and running initialization code
  • Subsequent invocations reuse the existing environment (warm start)
  • Cold starts can be mitigated with provisioned concurrency or by keeping the deployment package small

Layers​

  • A distribution mechanism for libraries, custom runtimes, or other dependencies
  • A function can use up to 5 layers at a time
  • Layers are extracted to the /opt directory in the execution environment
  • Useful for sharing common code across multiple functions
More Information

Pricing​

  • Charged based on the number of requests and the duration of execution
  • Duration is calculated from the time your code begins executing until it returns or terminates
  • Free tier: 1 million requests and 400,000 GB-seconds per month
More Information