Serverless Needs More Observability Tools
Developers of serverless applications have not changed a lot since we published the “Guide to Serverless Technologies” in 2018, which had surveyed over 300 users of serverless architecture. Back then, 78% primarily planned to use serverless architecture for new instead of existing workloads. Fast forward two years and 75% plan to build a greenfield serverless application in the next 12 months, according to a 2020 community survey managed by Jeremy Daly, who hosts the Serverless Chats podcast. The results’ raw data can be found in a GitHub repo.
The surveys are comparable, partly because we communicated with Jeremy during the question creation phase of the project. Then and now, the top use cases revolved around REST APIs and business logic. Then and now, excitement centered around design patterns rather than whiz-bang technologies. Now more than before, the breadth of adoption will continue to rise, but glaring obstacles are inhibiting enterprises from rapidly scaling up the workloads running on serverless architecture.
Debugging and monitoring continue to be the top serverless pain points. New design patterns and an increasing number of functions being managed are all related to these concerns. With over half of the respondents running 50+ functions in production, industry analyst Mark Wilcox noted, “The market for serverless monitoring and observability tools looks set for fantastic growth.”
In the past, users were willing to sacrifice portability between cloud environments and lack of control so they could make gains with speed of development and flexibility of scaling. Portability and control continue to create problems in the software development lifecycle, but fewer people think flexible scaling and increased speed are viewed as less likely outcomes. In 2020, event architecture is seen as the top area where serverless architecture is having a positive impact on the software development lifecycle.
According to container platform developer Karl Isenberg, this exemplifies a perpetual trade-off between platform abstractions — you can get control, flexibility and the ability solve specific issues, or you can choose an opinionated stack with a simple API and integrated user interface. By their very nature, AWS Lambda and other FaaS providers fall into the second category, but many enterprises are looking for more control, perhaps based on the Knative project.
Amazon Web Services is a sponsor of The New Stack.
Feature image by skeeze from Pixabay.
At this time, The New Stack does not allow comments directly on this website. We invite all readers who wish to discuss a story to visit us on Twitter or Facebook. We also welcome your news tips and feedback via email: firstname.lastname@example.org.