Datadog can now monitor serverless applications on Amazon Web Services’ (AWS) Lambda serverless platform. The new capability is enabled by an integration between Datadog’s hybrid cloud monitoring platform and AWS’ X-Ray distributed tracking and tracing system.
Serverless architectures, while growing in popularity, can be incredibly difficult to monitor. According to Datadog Product Manager Daniel Langer, Lambda, users are not exposed to the virtual machine (VM) that runs a Lambda function so traditional monitoring agents cannot access the VM. There is also a different set of monitoring information that you care about with serverless.
“Since the VM is abstracted away from the user, the user doesn’t care about traditional resource metrics like CPU load,” said Langer. Instead, users need metrics on errors, latency, and more, which can be obtained from a variety of sources. This includes cloud provider data metrics that are exposed from APIs, application and customer logs, and distributed tracing. “Being able to easily navigate between these three sets of data is currently challenging and takes developers a large amount of time and effort,” which is where Datadog comes in, he said.
AWS X-Ray provides users with end-to-end views of requests as they travel through an application to analyze and debug production and distributed applications. By integrating with Datadog, users can see serverless requests and executions to zero in on the source of errors and slowdowns. It also evaluates the performance of serverless functions and how they may impact overall application performance.
How this works is that when a request invokes multiple Lambda functions — that are connected by other AWS components — AWS X-Ray automatically instruments the functions and ties them into a single trace. Datadog has a trace detail view that combines the information into one place.
“Tracing is especially valuable for Lambda and other serverless platforms because it allows you to visualize how requests travel between the numerous components of a serverless architecture,” said Langer. Traces from Lambda are then combined with log and metric data from Datadog’s monitoring platform, which also tracks additional services, app components, and infrastructure areas.
Datadog combines this data into a unified view on its Cloud Functions UI. Using the interface, users can search, filter, and analyze Lambda functions and see Lambda metrics and logs with distributed request traces from the functions.
Datadog Synthetic Monitoring
Also this week, Datadog added a new capability to its unified monitoring platform that tracks application and API availability by simulating user traffic. Whereas users previously had to setup and manage the Datadog agent themselves, customers can now click a button to check URLs from different locations.
This Synthetics monitoring tool provides a way to “actively measure uptime, availability, and [the] response time of critical pages and transactions for multiple geographic locations,” said Albert Wang, a director of product management at Datadog.
In comparison to other providers that have similar capabilities, Wang said that Datadog brings together infrastructure metrics, application traces, and logs data into a single platform. Not only does this enhance the information compiled by the platform, but can make troubleshooting easier.