Easily build container-based AWS Lambda Functions

Pulumi makes it easy to take advantage of AWS Lambda’s ability to package serverless functions as container images. In this video, we’ll build a video thumbnailer. Let’s start with the bucket that we can push videos to. Before, running FFmpeg and Lambda was onerous. Now we can simply package it as part of our container, build it, and push the image to ECR. Pulumi makes it easy with a single line of code. Next, we need to give Lambda the right permissions for this function. We’re able to take advantage of the IDE, use automatic completion as well as TypeScript enums.

Let’s finish off our example by creating the Lambda function and referencing the image we just pushed. It’s easy to wire up the functions to the S3 bucket notifications in a single line of code. We export the bucket name for use outside of our program.

Finally, let’s augment our example by creating a callback to log information on when we create new thumbnails. That’s it. And just a couple of lines of code, we built the thumbnailing service. Let’s deploy it. Here, you can see it’s building the container image, pushing it as well as creating the various other resources we’ve defined. While that’s deploying, let’s take a quick look at our Docker file to see how we’re building our image. You can see it’s fairly straightforward. Let’s also take a quick look at the function handler. We’re simply shelling out to FFmpeg and copying the thumbnail to S3. Let’s grab a simple clip to try out. Our deployment finished in just a couple of minutes. Let’s watch the logs for our stack and try uploading our clip. We can see the logs for the function processing the clip. We can also see the logs for our other function that logs thumbnail creation. Let’s copy that thumbnail and take a look. Success. We hope you can see just how simple and easy Pulumi makes it to use container images with AWS Lambda. Our examples are in TypeScript, but Pulumi also supports Python, .NET and Go. Give it a try today.