(aws-cloudfront): few questions regarding Lambda@Edge function/stack

❓ General Issue

The Question

Hello, @njlynch.

I have a few questions regarding the new Lambda@Edge functionality.

1. When I try to create a L@E function, my stack is named *-us-west-2. Is this expected behavior? My understanding is that L@E stack will be deployed to us-east-1. My default region is us-west-2 when I run cdk doctor.

Successfully synthesized to /Users/robertd/code/cf/static-cloudfront-distributions/cdk.out
Supply a stack id (edge-lambda-stack-us-west-2, demo-appsStaticCFStack) to display its template.

1. If I create multiple L@E functions for the particular CF distro should I expect to see multiple CFN stacks for L@E functions? I see that everything is lumped under one stack right now.

2. I’m building a factory-style CDK application that will create multiple CloudFront distros with accompanying L@E functions (per distro) based on the configuration file. If these edge functions are all being lumped under the same stack, how do I keep them siloed (per CF distro) if their stacks are all named the same (i.e.: edge-lambda-stack-us-west-2).

Example (it doesn’t paint the whole picture but this is what I’m trying to do):

    const bucket = props.distroConfig.existingBucket ? s3.Bucket.fromBucketName(this, "existing-bucket", props.distroConfig.existingBucket) 
      : createBucket(this, props.distroConfig.stackName, this.account, logBucket);

    const edgeLambdas = props.distroConfig.lambdaEdgeConfig ? props.distroConfig.lambdaEdgeConfig.map((config, index) => {
      const edgeLambda = new cloudfront.experimental.EdgeFunction(this, `${props.distroConfig.stackName}-lambda-edge-${index}`, {
        runtime: config.runtime,
        handler: 'index.handler',
        code: lambda.Code.fromAsset(path.join(__dirname, 'lambda')),
        description: `Lambda Edge function for ${props.distroConfig.stackName}`,
      });
      
      return {
        functionVersion: edgeLambda.currentVersion,
        eventType: cloudfront.LambdaEdgeEventType.ORIGIN_REQUEST,
      }
    }) : undefined;

    const distro = new cloudfront.Distribution(this, "distro", {
      defaultBehavior: {
        origin: new origins.S3Origin(bucket),
        allowedMethods: cloudfront.AllowedMethods.ALLOW_ALL,
        // cachePolicy: cloudfront.CachePolicy.CACHING_DISABLED, // default: CACHING_OPTIMIZED 24hr
        cachedMethods: cloudfront.CachedMethods.CACHE_GET_HEAD_OPTIONS,
        viewerProtocolPolicy: cloudfront.ViewerProtocolPolicy.REDIRECT_TO_HTTPS,
        edgeLambdas,
      },
      certificate,
      comment: `${props.distroConfig.domainNames[0]} cloudfront distribution.`,
      logBucket: logBucket,
      logFilePrefix: `AWSLogs/${this.account}/cloudfront/${props.distroConfig.team}/static/${props.distroConfig.stackName}`,
      logIncludesCookies: true,
      defaultRootObject: "index.html",
      priceClass: cloudfront.PriceClass.PRICE_CLASS_100, // USA & EU
      domainNames: props.distroConfig.domainNames,
    });

Thanks! ✋

Environment

• CDK CLI Version: 1.82.0
• Node.js Version: 15.5.0
• OS: macOS Mojave
• Language (Version): TypeScript (4.1.3)

Other information

[njlynch]

Thanks for the feedback and questions! Questions 4 and 5 require a slightly deeper look; I expect we’ll end up with separate bug/feature requests for each, but let me do a quick probe into them first.

On the subjects of the L@E stack name:

1. When I try to create a L@E function, my stack is named *-us-west-2. Is this expected behavior? My understanding is that L@E stack will be deployed to us-east-1.
2. If I create multiple L@E functions for the particular CF distro should I expect to see multiple CFN stacks for L@E functions? I see that everything is lumped under one stack right now.
3. […] If these edge functions are all being lumped under the same stack, how do I keep them siloed (per CF distro) if their stacks are all named the same (i.e.: edge-lambda-stack-us-west-2).

Yes, the L@E stack will be deployed to us-east-1. The stack is named *-us-west-2, to indicate its origin. Your “primary”/CloudFront stack is in us-west-2, so the us-east-1 stack gets that name to tell you what region spawned the stack. This maybe wasn’t the clearest naming on my part. 😕

The current implementation was targeted at “simple” cases with a single CloudFront distro (per region), which is the most common app type I’ve seen; as such, I picked a naming convention I thought was simplest for the “base case”. However, in your case, this clearly won’t work terribly well. One solution might be a bug fix to have the stack named better (perhaps using the “parent” stack name as a seed); I’m open to suggestions here on what naming might make sense. The other approach for this case is #12136, which will allow you to explicitly specify the stack name for each L@E stack (this will go out with the next release).

[github-actions[bot]]

⚠️COMMENT VISIBILITY WARNING⚠️

Comments on closed issues are hard for our team to see. If you need more assistance, please either tag a team member or open a new issue that references this one. If you wish to keep having a conversation with other community members under this issue feel free to do so.