Datadog recently published the 'State of Serverless 2023' report, an annual analysis tracking the adoption of serverless technologies in cloud workloads. The major cloud providers are witnessing substantial growth in serverless usage, especially for container-based solutions.
As for previous years, Datadog used telemetry to determine which technologies their customers are using for each cloud provider, service, and language. The results suggest that the serverless ecosystem is experiencing continued growth, driven by organizations expanding their use of container-based applications within serverless deployments, with Google Cloud leading in the percentage of functions packaged as containers.
According to Datadog, the majority of organizations running workloads on AWS or Google Cloud now have at least one serverless deployment in place, with Azure following closely at 49 percent. Similarly, the latest CNCF Annual Survey reported growth of serverless architecture/FaaS to 53% of respondents.
Source: Datadog
Yan Cui, cloud consultant and AWS Serverless Hero, comments:
Paradigm shifts are hard and they always take longer to penetrate the mainstream thinking than we think. The same thing happened with FP and NoSQL. But the adoption is there (...)
Aside from the growth of container-based serverless platforms, the State of the Serverless 2023 report suggests that frontend development is an emerging category, with Vercel, Netlify, Cloudflare, and Fastly also providing serverless capabilities in this space. Jeremy Daly, CEO and founder of Ampt, comments in the Off-by-none newsletter:
I didn’t see any major surprises this year (...) Perhaps the most interesting fact was that Terraform dominates AWS Lambda deployments for larger organizations, though this also seems a bit obvious (due in large part to hybrid workloads). And finally, the fact that Frontend development is driving the creation of new serverless platforms just goes to show that cloud providers still haven’t figured out the DX for these services.
Marco Troisi, CTO at Trilo, agrees:
It's good and interesting as always. But there are no surprises there. It tells us nothing new. This is probably good news, as it means that serverless, despite the occasional bad press, is becoming more and more normalized and standardized.
For AWS Lambda, the most common type of serverless workload, Node.js, and Python, are still the dominant languages, with Java growing thanks to enterprise organizations finally adopting serverless technologies.
Source: Datadog
Even if Datadog collected data from over 20K companies, the report might not represent the full serverless community, as Zack Kanter, CEO @Stedi, highlighted on X (formerly Twitter). In the report, the company acknowledges:
While Datadog customers cover the spectrum of company size and industry (...) they skew toward adoption of cloud platforms and services more than the general population. All the results in this article are biased on the fact that the data comes from our customer base, a large but imperfect sample of the entire global market.
While the definition of serverless can extend to databases and other PaaS, Datadog focused on the adoption of serverless functions, managed container applications, and edge functions. For example, on AWS an organization is considered serverless when adopting Lambda, App Runner, Fargate, or CloudFront Functions.