Fat Lambda: Using a serverless strategy that exceeds the limits of Lambda

A serverless function is triggered by an event and runs on a platform that you do not need to manage. Each event fires the execution of a new serverless function, automatically scaling to the current demand. You only need to worry about your code (instead of servers) and you only pay for the resources you actually use, which typically carries very little cost. On AWS, serverless functions are typically written using the Lambda service. However, Lambda has certain limitations such as a run-time of 5 minutes and a 500 MB ephemeral disk. When you need to exceed those limits, for example if you have a long running query on a database that takes over 5 minutes, Lambda is not a good fit.

Enter Fargate. Introduced at re:invent 2017, Fargate is a serverless container service on AWS. Using Fargate, long running serverless solutions can be designed that break free of the limitations of Lambda. I like to call these solutions "Fat Lambdas." Fargate utilizes docker images to run tasks, so we have a lot of control over the environment and resources we are using, while still not having to worry about the actual server we are executing the code on. Only minimal docker knowledge is required to execute arbitrary python or other programs.

Please come and enjoy some free pizza and drinks and participate in a discussion led by Dan Rusk (VP of Software Development at Thorn Technologies) of the Fat Lambda serverless strategy, see a demo of the strategy in action, analyze the cost implications, and discuss how it can be applied in a couple situations where Lambda can't cut it.