Lambda concepts - Amazon Lambda
Services or capabilities described in Amazon Web Services documentation might vary by Region. To see the differences applicable to the China Regions, see Getting Started with Amazon Web Services in China (PDF).

Lambda concepts

Lambda runs instances of your function to process events. You can invoke your function directly using the Lambda API, or you can configure an Amazon service or resource to invoke your function.


A function is a resource that you can invoke to run your code in Lambda. A function has code to process the events that you pass into the function or that other Amazon services send to the function.


A trigger is a resource or configuration that invokes a Lambda function. Triggers include Amazon services that you can configure to invoke a function and event source mappings. An event source mapping is a resource in Lambda that reads items from a stream or queue and invokes a function. For more information, see Invoking Lambda functions and Using Amazon Lambda with other services.


An event is a JSON-formatted document that contains data for a Lambda function to process. The runtime converts the event to an object and passes it to your function code. When you invoke a function, you determine the structure and contents of the event.

Example custom event – weather data
{ "TemperatureK": 281, "WindKmh": -3, "HumidityPct": 0.55, "PressureHPa": 1020 }

When an Amazon service invokes your function, the service defines the shape of the event.

Example service event – Amazon SNS notification
{ "Records": [ { "Sns": { "Timestamp": "2019-01-02T12:45:07.000Z", "Signature": "tcc6faL2yUC6dgZdmrwh1Y4cGa/ebXEkAi6RibDsvpi+tE/1+82j...65r==", "MessageId": "95df01b4-ee98-5cb9-9903-4c221d41eb5e", "Message": "Hello from SNS!", ...

For more information about events from Amazon services, see Using Amazon Lambda with other services.

Execution environment

An execution environment provides a secure and isolated runtime environment for your Lambda function. An execution environment manages the processes and resources that are required to run the function. The execution environment provides lifecycle support for the function and for any extensions associated with your function.

For more information, see Lambda execution environment.

Instruction set architecture

The instruction set architecture determines the type of computer processor that Lambda uses to run the function. Lambda provides a choice of instruction set architectures:

  • arm64 – 64-bit ARM architecture, for the Amazon Graviton2 processor.

  • x86_64 – 64-bit x86 architecture, for x86-based processors.

For more information, see Lambda instruction set architectures (ARM/x86).

Deployment package

You deploy your Lambda function code using a deployment package. Lambda supports two types of deployment packages:

  • A .zip file archive that contains your function code and its dependencies. Lambda provides the operating system and runtime for your function.

  • A container image that is compatible with the Open Container Initiative (OCI) specification. You add your function code and dependencies to the image. You must also include the operating system and a Lambda runtime.

For more information, see Lambda deployment packages.


The runtime provides a language-specific environment that runs in an execution environment. The runtime relays invocation events, context information, and responses between Lambda and the function. You can use runtimes that Lambda provides, or build your own. If you package your code as a .zip file archive, you must configure your function to use a runtime that matches your programming language. For a container image, you include the runtime when you build the image.

For more information, see Lambda runtimes.


A Lambda layer is a .zip file archive that can contain additional code or other content. A layer can contain libraries, a custom runtime, data, or configuration files.

Layers provide a convenient way to package libraries and other dependencies that you can use with your Lambda functions. Using layers reduces the size of uploaded deployment archives and makes it faster to deploy your code. Layers also promote code sharing and separation of responsibilities so that you can iterate faster on writing business logic.

You can include up to five layers per function. Layers count towards the standard Lambda deployment size quotas. When you include a layer in a function, the contents are extracted to the /opt directory in the execution environment.

By default, the layers that you create are private to your Amazon account. You can choose to share a layer with other accounts or to make the layer public. If your functions consume a layer that a different account published, your functions can continue to use the layer version after it has been deleted, or after your permission to access the layer is revoked. However, you cannot create a new function or update functions using a deleted layer version.

Functions deployed as a container image do not use layers. Instead, you package your preferred runtime, libraries, and other dependencies into the container image when you build the image.

For more information, see Working with Lambda layers.


Lambda extensions enable you to augment your functions. For example, you can use extensions to integrate your functions with your preferred monitoring, observability, security, and governance tools. You can choose from a broad set of tools that Amazon Lambda Partners provides, or you can create your own Lambda extensions.

An internal extension runs in the runtime process and shares the same lifecycle as the runtime. An external extension runs as a separate process in the execution environment. The external extension is initialized before the function is invoked, runs in parallel with the function's runtime, and continues to run after the function invocation is complete.

For more information, see Lambda extensions.


Concurrency is the number of requests that your function is serving at any given time. When your function is invoked, Lambda provisions an instance of it to process the event. When the function code finishes running, it can handle another request. If the function is invoked again while a request is still being processed, another instance is provisioned, increasing the function's concurrency.

Concurrency is subject to quotas at the Amazon Region level. You can configure individual functions to limit their concurrency, or to enable them to reach a specific level of concurrency. For more information, see Configuring reserved concurrency.


When you invoke or view a function, you can include a qualifier to specify a version or alias. A version is an immutable snapshot of a function's code and configuration that has a numerical qualifier. For example, my-function:1. An alias is a pointer to a version that you can update to map to a different version, or split traffic between two versions. For example, my-function:BLUE. You can use versions and aliases together to provide a stable interface for clients to invoke your function.

For more information, see Lambda function versions.


A destination is an Amazon resource where Lambda can send events from an asynchronous invocation. You can configure a destination for events that fail processing. Some services also support a destination for events that are successfully processed.

For more information, see Configuring destinations for asynchronous invocation.