AWS DevOps Blog

Build ARM-based applications using CodeBuild

AWS CodeBuild has announced support for ARM-based workloads, which will allow you to build and test your software updates natively, without needing to emulate or cross-compile. ARM is a quickly growing platform for application development today and if you rely on emulation and/or cross-compile, the downside is time and reliability. However, a more native approach can be faster and more reliable: Enter ARM-based workload support.

In this post, you will learn how to build a sample Java application with an ARM-based Docker image, you will then upload the artifact to an S3 bucket.


A new repository in CodeCommit with the code from the sample Java application linked above has already been created. A working knowledge of git and how to fork or clone within your source provider is a pre-requisite.

Configuration Steps

Working with our source code:

  1. Fork or clone the repo and upload/push the code to your source provider of choice. As of this writing, CodeBuild supports the following Source Providers: S3, CodeCommit, BitBucket, GitHub, and GitHub Enterprise.
  2. Go to your IDE of choice* and within your repo/source create a new file named buildspec.yml and copy in the following code. *In this post, the AWS Cloud9 IDE will be referenced when discussing edits. (buildspec.yml reference page)
    version: 0.2
               java: corretto8
                - echo Starting Java build at `date`
                - mvn package
                - echo Finished build of Java Sample at `date`
            - 'target/aws-java-sample-1.0.jar'

    For this sample, specify that your container run an Amazon Corretto 8 Java environment. An artifact file will also be output, which we will be sent to an S3 bucket later in the process.

    The two phases are:

    1. Install – Since version 0.2 is being used, the Install phase is required to specify the runtime-version.
    2. Build – This phase is where the commands used to build the software will be passed.
  3. Once the buildspec.yml file has been added and saved, you will commit your changes and push your code to your source.
    We are doing a git push to our repo.


Creating your CodeBuild Project:

Now that you have created your source code, it’s time to create your CodeBuild project. For this post, The AWS Management Console will be used, though other tools such as AWS Cloud Dvelopement Kit (CDK), AWS CloudFormation, or the AWS CLI can also be leveraged.

Creating your artifact destination:

The first thing you are going to do is create where your artifact will be stored. For this blog, you are going to put your artifact into an S3 bucket.

1) In the console search bar type ‘S3.’

2) Select ‘S3’ to go to the S3 Console

image displaying how to search for the S3 service in the AWS Management Console

3) Select ‘Create bucket’ from the top left of the console.

image showing where to click to create your S3 Bucket


4) Type in your bucket name and Region and click ‘Next.’ For this blog, you are going to use the name “mydemobuildbucket.” It is important to note that it must an be all lowercase and globally unique name.

image showing the fields in the name and region of the S3 bucket creation.

5) Leave the defaults as is for the configuration page and select ‘Next.’

image displaying to leave the configuration options as they are and click next.

6) Under the permissions tab, choose to ‘Block all public access’ to the bucket, then click ‘Next.’ This well help keep your artifact secure.

image displaying to select block all public access and select next. this is important to secure our bucket.

7) The Review pane is where you can verify all of your settings. Once you have confirmed that all settings are correct, click ‘Create bucket’ to finish.

image showing the summary of the bucket creation process and to click create bucket.

You should see your S3 Bucket. With your bucket, you can now create your build project.


Create CodeBuild Project:

If you have worked with CodeBuild in the past, most of this will look familiar, however, as part of the ARM release, a few new options have been added within the Build Environment section that will let you build and test any of your ARM-based applications.

1) Go to the console search and type ‘CodeBuild’ and select the service.

image showing how to search for the CodeBuild service in the AWS Management Console

2) In the top right corner click ‘Create build project.’

Image showing where to click to create a build project

3) Enter a name for your project. (‘myDemoBuild’ will be used as the default name in this post.) Descriptions are optional but can be useful.

image showing entering our project name myDemoBuild and an optional description

4) Select your source provider. I am going to use CodeCommit and select my repo and the branch where my new code is located.

image shows selecing our Source provider, Repository name, reference type, and branch name.

5) This is where the differences are that I mentioned earlier. We are now going to setup our Build Environment. For ARM support we must select the following options:

  1. Operating System: Amazon Linux 2 (At time of publishing, ARM support only supports using the Amazon Linux 2 operating system.)
  2. Runtime(s): Standard
  3. Image: amazonlinux-aarch64-standard

You can either create a new service role or select an existing service role if you have previously created one. I am going to create a new role called myDemoRole. The system will automatically create the required permissions to allow CodeBuild to access the resources based on your input from the Project. In a production environment, I would recommend creating a service role that follows a least access needed principal, instead of creating a holistically new role.

image showing the environment settings for CodeBuild.

6) Configure buildspec settings. I am going to select ‘Use a buildspec file’ and leave the name blank as it will default to buildspec.yml. However, you can select a specific name if you have multiple buildspec files for use with difference environments. E.g., buildspec-prod.yml, buildspec-staging.yml, buildspec-dev.yml

Image showing us select to use a buildspec file.

7) Configure your artifact settings. I am going to upload my artifact to the S3 Bucket we created earlier. I have named the file for simplicity, but any name can be used. I have selected to Zip the artifact file, however, that is not required.

image showing the artifact configuration for our CodeBuild project.

8) Configure logging. I am going to enable CloudWatch logging so that the build logs are uploaded to the logging service.

image showing configuring logs which is an optional step.

9) Select ‘Create build project.’

image showing the review panel where you can review all of the settings for your CodeBuild Project.

Run the build:

Now that your application is created, your S3 bucket has been built, and you’ve created your build project, it is time to run your build. If everything is successful, your artifact file should be stored in your S3 bucket.

1) After you have created your build project, you should now be in the build page which allows you to run/edit/delete your build project. Select ‘Start build’ in the top right part of the page.

image showing how to start the build you just created.

2) You can review or override the build settings that you configured when setting up the project. Once you have verified settings, click ‘Start build’ in the top right.

image showing the verify run options and to kick off the build.

3) Once your build has been started, it will run through the steps of you buildspec file (this can take a while depending on your application). Once complete, the status should show “Succeeded.”

image showing where to check for build status.

If for any reason your build did not succeed, look in the phase details to find the error. Validate all of your settings are correct and use the documentation to help you troubleshoot any issues.

4) Now that your build has been successful, verify that your artifact is in the S3 bucket you created.

image showing where you will find the Artifact in the S3 Bucket.


Clean Up

Reminder, if you created any resources just for testing purposes, you should delete them to keep from incurring additional cost.

Make sure and check the following when cleaning up:

  • S3 bucket
  • CodeBuild Project
  • CodeCommit repository
  • Cloud9 Environment


We have walked through the process of using CodeBuild to build a sample Java application in the new ARM environment. Now that you have built your ARM-based artifact, you can download it for any local use or get started developing your own ARM-based applications using AWS Developer Tools.