A recent survey of Amazon Web Services customers provides a general profile for the typical AWS Lambda customer — namely, a large enterprise using Python and containers.
This is the response to a report called “The State of Serverless”, released by Datadog, a provider of cloud asset tracking solutions for DevOps teams. Datadog analyzed the AWS usage data for its “thousands”, to determine when, where, and why they use Lambda. This is AWS’ serverless computing platform, which first launched in late 2014. According to Datadog’s methodology notes, a customer is considered a Lambda user if it “ran at most five distinct Lambda functions per month.”
Lambda effectively frees code running from the need for maintaining physical infrastructure. It is event-driven, which allows organizations to set parameters that will trigger Lambda to run the code they have written.
Datadog’s findings show that Lambda is “the most mature, widely used serverless platform in our userbase and beyond.” It was used by almost 50 percent of AWS users in Datadog’s customer pool in January 2020. This is an increase of 35 percent from January 2019.
Lambda is also more commonly used by enterprises (about 80%), than “small,” medium, or large organizations. Datadog found that Lambda is more common in larger organizations than it is in smaller ones.
The report stated that Lambda is not limited to cloud-native early users or niche use cases. “Serverless functions are now widespread in use across a variety companies with AWS infrastructure footprints.”
Another interesting finding is that Python and Node.js were the most used languages for Lambda functions. Python accounts for 47% of all functions, while Node.js is responsible for 39%. Java is in third place, with less than 1%. Datadog attributes the dominance and popularity of Python and Node.js both to their presence in the Lambda service as well as their popularity.
The report stated that Lambda was launched in preview by AWS in 2014 using Node.js, the first supported runtime. It was then extended to Java and Python support in 2015. “Support for C#, Go, and Ruby (via.NET Core) were added in 2018”
The report also found that Lambda is a heavy product among container users. Around 80 percent of companies that use containers are also Lambda adopters. Datadog says there is a good reason for this correlation.
Although serverless functions are different from containers, they are often embraced for similar reasons. For example, to reduce infrastructure concerns and make operations more efficient. While Lambda and container infrastructure can be used in some cases (e.g. using Lambda functions that trigger Amazon Elastic Container Service tasks), many organizations may run them separately to meet different needs. A company might run their entire application in a container cluster while offloading short-running, bursty tasks (such payment processing) to serverless services.
Datadog has the complete report on Lambda usage trends.