By using AWS re:Post, you agree to the Terms of Use
AWS re:Postre:Post
/Lambda source code updating issue during its container image building via cdk/

Lambda source code updating issue during its container image building via cdk

0

I am making a lambda function through DockerImageFunction with DockerImageCode.fromImageAsset() from aws-cdk-lib/aws-lambda. The Dockerfile contains a collection of codec tools, and the lambda handler function code. Moreover, those codec tools takes 5 min to build, and hardly ever needs to be updated. The problem is that there are lot of lambda functions use the same codec tools, and the same codec tools get rebuilded everytime my lambda handler function code has been commited. Is there a way to craete those codec tools first, and for those lambda functions which need the tools during image building, just import it? Thanks!

Topics
ServerlessComputeDeveloper ToolsDevOps
Tags
AWS LambdaDeveloper ToolsDevOps
strange-folk-df787
asked a month ago51 views
1 Answers
0
Accepted Answer

You could create a base image that contains all your shared dependencies (i.e. codec tools) and then reference this base image in the Dockerfile on your lambda functions. The base image should be built as a pre-requisite step and pushed to an ECR repository.

ogaumond
answered a month ago
  • strange-folk-df787 a month ago

    Would you mind pointing out how exactly to create a base image and then reference it in the lambda? I was making the base image by the DockerImageAsset from aws-cdk-lib/aws-ecr-assets, and then using asset.imageUri to reference it when creating a lambda, like the following:

    new DockerImageFunction(this, "lambda-handler", {code: DockerImageCode.fromImageAsset(path.join("/dockerfile/path/"), {buildArgs: { CODEC_SOURCE: asset.imageUri} })})

    But I got an Error said:Cannot use tokens in keys or values of "buildArgs" since they are needed before deployment.

    Thanks!

  • ogaumond a month ago

    The error you get from CDK is because the imageUri is only known at execution time, but the buildArgs are needed at synth time. You may need to separate the build of the base image in a separate process to have the base image pushed to a known ECR registry that can then be referenced from the lambda stack. The following construct could be used to publish the base image to ECR: https://github.com/cdklabs/cdk-ecr-deployment

You are not logged in. Log in to post an answer.

A good answer clearly answers the question and provides constructive feedback and encourages professional growth in the question asker.

Guidelines for Answering Questions

Relevant questions