Lambda Containers with Rails; A Perfect Match!

A few weeks ago when I started this serverless & container series we talked about how to use Docker in your AWS Lambda projects with SAM. Who could have known that today I get to share in the recent AWS Lambda Container Image Support news by announcing that Lamby works out of the box with both ZIP and Container deployment packages!

How & Why?

From a Rails perspective with Lamby, nothing really changes. Lamby is a simple Rack adapter to mount your Rails application within any chosen Lambda integration point, most commonly, API Gateway. However, we did rewrite our quick start guide to use containers by default. Take it for a test drive to see your own Rails application on AWS Lambda!

From a Lambda perspective, using containers is a "simple" deployment strategy option. All the same guarantees for the ZIP strategy apply. But this news also brought a huge buff for Lambda as well.

• Memory: 128 MB to 10,240 MB, in 1-MB increments.
• vCPU: Up to 6 cores using memory setting.
• Container: 10 GB code package size.

Rather than duplicate work on the basics of Lambda functions as container images, I highly recommend these two posts:

• Using container images with AWS Lambda - Michael Hart breaks things down to their smallest components using a simple Go PDF converter microservice.
• Packaging AWS Lambda functions as container images - A guest post by James Beswick on A Cloud Guru.

Much More Coming!

So re:Invent is not over and I'm sure AWS has a lot more cool things coming. From me, expect to see some very exciting new tools over the next few weeks.

Up next, Lambdakiq! A 100% drop in replacement for Sidekiq that leverages AWS managed services like SQS. Never again worry about side car containers, worker instances, process management for polling, or threading & concurrency. All these tools and worries evaporate when you integrate Rails with AWS! Stay tuned.

Rough Edges

First, I'd like to congratulate all the AWS teams that have been working their butts off to bring us these awesome features. I can only imagine the quality & technical scale those teams operate under. Much respect.

While working with AWS SAM for the Lamby quick start I did encounter a few issues. Some of which may be me. Documenting them here just in case.

• Using ImageConfig in SAM can cause a SUPERNOVA
• Can SAM & Lambda teams both leverage ECR/Docker repositories?
• Can SAM's Makefile be used with Containers?
• Does SAM require an S3 Bucket for Containers?
• Can SAM's also create your ECR repository?