OpenTelemetry: What’s New With the Second-Biggest CNCF Project?
Morgan McLean, a former employee at Microsoft and Google who’s now at Splunk, has long wrestled with solving the challenges of observability in large-scale systems.
As a product manager early in his career, he would sometimes write internal tools. “One of my biggest points of frustration was always, when we were working on large, high-scale services, being able to actually debug them when things went wrong,” said McLean, co-founder of OpenTelemetry and senior director of product management at Splunk, a Cisco company, in this episode of The New Stack Makers.
In those days, he said, “when we were on call or when we were pushing out rollouts, if there was any risk of instability, it would take a nontrivial amount of time to chase down what went wrong, and that was just due to our tools.”
In this episode, McLean talked to Alex Williams, TNS founder and publisher, about the past, present and future of OpenTelemetry, the open source framework that helps software engineers collect and analyze data about how their systems and applications are performing.
OpenTelemetry was created from the merger of two projects in 2019: OpenTracing and OpenCensus. It’s a Cloud Native Computing Foundation incubating project, and it’s been rapidly adopted as a part of many organizations’ observability strategies.
OTel, McLean said, emerged to solve the problems of scale that emerged as organizations began to embrace Kubernetes. “It meant now we had to spend a lot of time extracting data from customers’ applications in order to make those tools work. It sounds relatively trivial, right? You’re just switching platforms. What’s the big deal?
But, he added, “because things like OpenTelemetry didn’t exist, it’s actually a heck of a lot of work to be able to get distributed traces, to get application metrics, to get various other types of data out of customers’ environments. Because it means you need to integrate with every language runtime, every framework that customers use, and every database that they’re using.”
Integrating with all of them, he concluded, is “untenable for any one person, organization, whatever, to go and maintain that.” A new set of standards was called for. Hence the subsequent creation of OpenTracing (begun with contributions by LightStep and Uber) and OpenCensus (started at Google) — and eventually, OpenTelemetry.
Alleviating ‘Points of Frustration’ With OTel
OpenTelemetry is now the second most active open source project‚ after Kubernetes, in the CNCF, McLean told the Makers audience, with more than 1,200 developers checking in code to OTel repos per month.
“It’s very obvious to me now, several years in, the impact that this has on the industry,” he said. “People can use these tools. Developers everywhere can gain these deep insights into their applications and their infrastructure as a result.”
But plenty of work needs to be done. For instance, documentation could be improved, McLean said. “Open source project management — in terms of how the organization is structured and who’s working on what and things like that — is not really radically different than running a project in a company or really in any kind of environment. It’s all just human project management.”
OTel innovations loom on the horizon. For instance, McLean spoke to TNS at KubeCon + CloudNativeCon North America last November about profiling signals, slated to move into general availability in 2025.
Profiling, he said, is the fourth major observability signal — alongside logs, tracing and metrics.
“With profiling, you’re actually getting insight into the performance a level deeper into the application itself, and the actual functions or methods inside of that application. You can see their memory consumption. You can see how they call each other. You can see their CPU consumption.”
McLean also told the Makers audience about efforts made over the past couple of years, both at Splunk and within the OTel community, on the OTel operator and other efforts to create components to make adoption easier for the end user.
“As you see more and more improvements get made to open telemetry that just make it automatic, or as it gets built into more platforms and run times just out of the box, you’re going to see — I mean, it’s already rapid adoption — but you will see that accelerate even further. The few points of frustration that people may have had with it evaporate.”
Check out the full episode to learn more about what’s new and what’s ahead for OpenTelemetry.
