AppRunnerService always failed upon CDK update

11:03:22 PM | UPDATE_FAILED        | AWS::AppRunner::Service                         | AppRunnerDC2BB513
Internal Failure

 ❌  app failed: Error: The stack named app failed to deploy: UPDATE_ROLLBACK_COMPLETE: Internal Failure

The AppRunner Service is always failed internally, causing deploy failure on my CDK stack update on a deployed stack.

const imageAsset = new ecrAssets.DockerImageAsset(this, 'ImageAsset', {
    directory: path.join(__dirname, 'docker'),
});
const appRunner = new apprunner.Service(this, 'AppRunner', {
    source: apprunner.Source.fromAsset({
        imageConfiguration: {
            port: 8888,
            environment: {
                ... // Some other CDK construct output
                version: process.env.npm_package_version ?? 'Undefined',
            },
        },
        asset: imageAsset,
    }),
    cpu: Cpu.ONE_VCPU,
    memory: Memory.FOUR_GB,
});

My CDK stack also has other constructs such as Lambda, S3, etc.

Often times the docker image remains untouched, but environment variable (e.g. version) is changed.

I have checked the CloudWatch logs of AppRunner, and cannot found any errors on deployment and execution.

I have to destroy the stack and re-deploy it whenever this error happens, which means the output paths will be different.