Datadog Report Shows More Reliance on Serverless Computing
A report published this week by Datadog, a provider of a monitoring and observability cloud service, showed that serverless computing is now widely employed across all the major cloud computing platforms.
An analysis of serverless computing use across the Datadog customer base found more than 70% of Amazon Web Services (AWS) and 60% of Google Cloud customers currently use one or more serverless solutions, followed by 49% of Microsoft Azure customers.
Danny Driscoll, a Datadog product manager, said the primary reason is that many of those organizations have already developed container applications and DevOps workflows that they are extending to run workloads on serverless computing frameworks by invoking a microservice.
For example, 66% of the serverless workloads running on Google Cloud are container-based, compared to 26% for AWS and 22% for Azure. Google has a much higher percentage because it makes a specific Cloud Run service for container applications available, noted Driscoll.
Most of the use cases that employ serverless frameworks involve asynchronous event-driven applications that are using serverless frameworks to execute a specific function. That approach reduces the overall size of the application when specific tasks are executed.
The report also noted that Python and Node.js are still the most popular languages used among developers using the AWS Lambda serverless platform. In fact, well over half of the invocations in Datadog’s Lambda dataset were from functions written in Python or Node.js. Java is the third most common Lambda language, closely followed by custom runtimes and Go. Together, they accounted for just under a quarter of all the Lambda invocations. Custom runtimes are the fastest-growing types of functions, however, having increased 50% in the last year on the AWS Lambda service, the report said.
In addition, the Datadog report noted that IT teams are dealing with cold start issues that can adversely impact containers by increasing the amount of memory allocated to a Lambda function to reduce cold start durations.
Finally, the Datadog report also noted that other serverless computing platforms are also now starting to gain traction. Modern frontend development platforms and content delivery network (CDN) platforms such as Vercel, Netlify, Cloudflare and Fastly all now provide access to serverless frameworks. And 7% of all organizations using Datadog to monitor serverless workloads in a major cloud also run workloads using at least one of these emerging cloud platforms. Of these, 62% use a frontend development platform such as Vercel or Netlify, while 39% use CDNs offering edge compute capabilities such as Cloudflare or Fastly.
It’s hard to say what percentage of the platforms that DevOps teams will manage will be serverless computing frameworks, but the Datadog report makes it clear they are here to stay. The challenge is that these applications are highly dynamic, so there is an acute need for monitoring and observability tools. Regardless of approach, however, developers will continue to invoke a mix of these platforms, regardless of how challenging they are for DevOps teams to manage.
