AWS Lambda Custom Runtime Function Project

Prerequistes

• AWS Account – You need to have a valid AWS account to get started with it. Navigate to https://console.aws.amazon.com/ to get started with Amazon Web Services
• Visual Studio Code – Download and install Visual Studio Code in your machine. VS Code is available for Windows, Linux, and Mac
• Code Development SDK – You need to install the relevant SDK that you are going to use to code your application. This can be either .NET SDK, Node JS SDK or the Python SDK
• AWS SAM CLI – This is a command line utility from Amazon that enables us to develop and test our serverless applications locally. Although it is not mandatory for the toolkit, but still it is recommended to install it on the local machine

AWS Extensions for .NET CLI

This sample project has been built with this tool: https://github.com/aws/aws-extensions-for-dotnet-cli

To install these tools use the dotnet tool install command.

    dotnet tool install -g Amazon.Lambda.Tools

To update to the latest version of one of these tools use the dotnet tool update command.

    dotnet tool update -g Amazon.Lambda.Tools

Project structure

The source code of this starter project consists of:

• Function.cs - contains a class with a Main method that starts the bootstrap, and a single function handler method
• aws-lambda-tools-defaults.json - default argument settings for use with Visual Studio and command line deployment tools for AWS

The generated Main method is the entry point for the function's process. The main method wraps the function handler in a wrapper that the bootstrap can work with. Then it instantiates the bootstrap and sets it up to call the function handler each time the AWS Lambda function is invoked. After the set up the bootstrap is started.

The generated function handler is a simple method accepting a string argument that returns the uppercase equivalent of the input string. TODO: Replace the body of this method, and parameters, to suit your needs.

There is also a test project.

Visual Studio

Deploying and invoking custom runtime functions is not yet available in Visual Studio

Steps to follow to get started from the command line

• Once you have edited your template and code you can deploy your application using the Amazon.Lambda.Tools Global Tool from the command line. Version 3.1.4 or later is required to deploy this project.

• Install Amazon.Lambda.Tools Global Tools if not already installed.

      dotnet tool install -g Amazon.Lambda.Tools
  

• If already installed, check if new version is available.

      dotnet tool update -g Amazon.Lambda.Tools
  

• Create the project naming it LambdaNet6

      dotnet new lambda.CustomRuntimeFunction --name LambdaNet6
  

Manage the created project

1. Deploy function to AWS Lambda

   It deploys the .NET Core Lambda project directly to the AWS Lambda service. The function is created if this is the first deployment. If the Lambda function already exists then the function code is updated. If any of the function configuration properties specified on the command line are different, the existing function configuration is updated.

       cd "LambdaNet6/src/LambdaNet6"
       dotnet lambda deploy-function LambdaNet6
   

   Note: To avoid accidental function configuration changes during a redeployment, only default values explicitly set on the command line are used. The defaults file is not used.

   The output details will be very similar to:

       Project Home: https://github.com/aws/aws-extensions-for-dotnet-cli, https://github.com/aws/aws-lambda-dotnet
   
       Enter AWS Region: (The region to connect to AWS services, if not set region will be detected from the environment.)
       eu-central-1
       Executing publish command
       ... invoking 'dotnet publish', working folder 'c:\Code\serverless.net60\LambdaNet6\src\LambdaNet6\bin\Release\net6.0\publish'
       ... dotnet publish --output "c:\Code\serverless.net60\LambdaNet6\src\LambdaNet6\bin\Release\net6.0\publish" --configuration "Release" --framework "net6.0" --self-contained true /p:GenerateRuntimeConfigurationFiles=true --runtime linux-x64
       ... publish: Microsoft (R) Build Engine version 17.0.0+c9eb9dd64 for .NET
       ... publish: Copyright (C) Microsoft Corporation. All rights reserved.
       ... publish:   Determining projects to restore...
       ... publish:   Restored c:\Code\serverless.net60\LambdaNet6\src\LambdaNet6\LambdaNet6.csproj (in 10,23 sec).
       ... publish:   LambdaNet6 -> c:\Code\serverless.net60\LambdaNet6\src\LambdaNet6\bin\Release\net6.0\linux-x64\bootstrap.dll
       ... publish:   LambdaNet6 -> c:\Code\serverless.net60\LambdaNet6\src\LambdaNet6\bin\Release\net6.0\publish\
       Zipping publish folder c:\Code\serverless.net60\LambdaNet6\src\LambdaNet6\bin\Release\net6.0\publish to c:\Code\serverless.net60\LambdaNet6\src\LambdaNet6\bin\Release\net6.0\LambdaNet6.zip
       ... zipping: Amazon.Lambda.Core.dll
       ... zipping: Amazon.Lambda.RuntimeSupport.dll
       ... zipping: Amazon.Lambda.Serialization.SystemTextJson.dll
       ... zipping: bootstrap
       ... zipping: bootstrap.deps.json
   

2. Invoke Function It invokes the Lambda function in AWS Lambda passing in the value of --payload as the input parameter to the Lambda function.

       dotnet lambda invoke-function LambdaNet6 --payload "hello lamBDA"
   

   The string is not correctly updated, being now uppercase.

       Amazon Lambda Tools for .NET Core applications (5.2.0)
       Project Home: https://github.com/aws/aws-extensions-for-dotnet-cli, https://github.com/aws/aws-lambda-dotnet
   
       Payload:
       "Architecture: X64, .NET Version: 6.0.0 -- HELLO LAMBDA"
   
       Log Tail:
       START RequestId: 6cbe418a-cbed-4e57-86c6-1bb4c853181b Version: $LATEST
       END RequestId: 6cbe418a-cbed-4e57-86c6-1bb4c853181b
       REPORT RequestId: 6cbe418a-cbed-4e57-86c6-1bb4c853181b  Duration: 2.14 ms       Billed Duration: 3 ms   Memory Size: 256 MB     Max Memory Used: 68 MB
   

3. Execute unit tests

       cd "LambdaNet6/test/LambdaNet6.Tests"
       dotnet test
   

   The output will be:

       Starting test execution, please wait...
       A total of 1 test files matched the specified pattern.
   
       Passed!  - Failed:     0, Passed:     1, Skipped:     0, Total:     1, Duration: < 1 ms - LambdaNet6.Tests.dll (net6.0)
   

Arm64

If you want to run your Lambda on an Arm64 processor, all you need is to do is add "function-architecture": "arm64" to the aws-lambda-tools-defaults.json file. Then deploy as described above.

Improve Cold Start

In the csproj file the PublishTrimmed and PublishReadyToRun properties have been enable to optimize the package bundle to improve cold start performance.

PublishTrimmed tells the compiler to remove code from the assemblies that is not being used to reduce the deployment bundle size. This requires additional testing to make sure the .NET compiler does not remove any code that is actually used. For further information about trimming check out the .NET documentation: https://docs.microsoft.com/en-us/dotnet/core/deploying/trimming/trimming-options

PublishReadyToRun tells the compiler to compile the .NET assemblies for a specific runtime environment. For Lambda's case that means Linux x64 or arm64. This reduces the work the JIT compiler does at runtime to compile for the specific runtime environment and helps reduce cold start time.