Debugging Tip For Node on AWS Lambda Edge

Lambda is hands down my favorite AWS service. The capability to run serverless functions on demand and at scale is very powerful. A more recent capability of Lambda is that it can be run in response to certain Cloud Front events. AWS Cloud Front is a content delivery network. This integration of Lambda is referred to as Lambda Edge. You can use Lambda Edge the same way that you use normal Lambda. However, as a trigger you select Cloud Front and select the associated distribution and events to use as triggers.

When you use Cloud Front as a trigger, you can then access the option within your latest Lambda function to deploy to Lambda@Edge. This option will deploy your Lambda function across the Cloud Front edge servers kind of like a Cloud Front invalidation.

This is where it becomes a little more complex. Thankfully the same code is deployed to all edge servers.

The biggest tip is to use CloudWatch logs to see if your functions are running. The big catch here is that your CloudWatch logs will be within the region of the originating requests. There are a LOT of regions. Thankfully you can do simple logging in Lambda by typical language console output methods. For Node Console.log will do the job. For Java System.out.println and the like will get output into your CloudWatch logs.

Also, watch out for lagging code deployment. As of August, 2019 I have experienced lagging edge deployments when using "Deploy to Lambda@Edge". I waited a day and the correct code was magically running.