OpenTelemetry at Grafana Labs: the latest on how we're investing in the emerging industry standard

Here at Grafana Labs, open source has always been core to what we do. So it should come as no surprise that we’re going all in on OpenTelemetry—an open source project that’s quickly becoming an industry standard for vendor-neutral telemetry. 

As Grafana Labs CTO Tom Wilkie put it at GrafanaCON 2025, Grafana Labs has been disrupting the observability ecosystem for more than a decade with our big tent philosophy. It’s a philosophy that prioritizes interoperability, starting from our earliest days with visualizations in Grafana to our ongoing contributions to Prometheus and our many other open source projects. This has helped commoditize the market so you can store more data and get better insights at a lower cost. And the market is once again being disrupted—this time by OpenTelemetry

“Users now really only have to instrument once and that instrumentation can be sent to any technology out there,” he said in the GrafanaCON keynote. “I’m really, really proud to have been part of this movement to kind of disrupt what I saw as a relatively stale ecosystem of proprietary technology with this really cool open source.”

In this blog, we’ll take a moment to reflect on what’s top of mind for us when it comes to OpenTelemetry. We’ll share some of our key takeaways from GrafanaCON 2025, our biggest open source event of the year. We’ll also provide an update on what we’ve been doing to increase our OpenTelemetry contributions and what we’re most excited about for the months ahead.

Why OpenTelemetry is so important to observability—and Grafana labs

Developer Program Director Ted Young shared several key reasons why OpenTelemetry is important during the GrafanaCON 2025 keynote. For starters, he touched on the role of vendor neutrality and the need for wider adoption across the industry.

“When you’re standing up observability, one of the heaviest lifts is adding all the instrumentation to all of your applications, services and infrastructure,” said Young, co-founder of OpenTelemetry project and a relatively new addition to the Grafana Labs team. “If you have to rip and replace all of that instrumentation every single time you want to try a different observability tool, we don’t think that’s really acceptable. We think you should write once, run everywhere, and we believe that OpenTelemetry is the tool that provides that feature.” 

Another important aspect of OpenTelemetry is that it standardizes the data coming from your system through Semantic Conventions, which provide predictability that’s really helpful when it comes to any downstream tasks, whether that’s building dashboards or using AI to find correlations.

“What it basically means is that if we’re reporting an HTTP request, it’s always reported the same way,” he said. “Whether it’s coming from a Java service or Python application or Kubernetes, HTTP should always look like HTTP. Same thing for SQL and everything else. It should be very, very regularized.”

He also discussed the importance of unifying different signals.

“Traditionally tracing, metrics, and logs, and now profiling, were all separate tools and they had completely separate stacks and they were siloed from each other that made it hard to move across all of these different data systems,” he said. “OpenTelemetry doesn’t just take those different signals and stack them next to each other. It actually integrates all of that information into a single graph.”

Young went on to talk about some of our recent contributions to the OpenTelemetry project. We launched Grafana Beyla, an open source eBPF-based, zero-code instrumentation tool, in 2023, and we recently donated the project upstream under the new project name OpenTelemetry eBPF Instrumentation. (Note: Beyla will continue to exist as Grafana Labs’ distribution of the OpenTelemetry eBPF Instrumentation tool.)

“We’re excited to have this as part of OpenTelemetry because standard data, we’d like it to be the case where whether it’s auto instrumentation or manual instrumentation, the data looks the same,” Young said. “And we’re also interested in unified telemetry—in particular combining eBPF with profiling.”

How we’re building OpenTelemetry into our tools with Grafana Alloy and more

Young returned to the stage later that day with Senior Software Engineer Johanna Öjeling, where they talked about Grafana Alloy. The open source distribution of the OpenTelemetry Collector is focused on bringing together the best of OTel and Prometheus for a truly optimized hybrid experience.

The pair discussed the history and definition of the OpenTelemetry Collector, and how Grafana Alloy fits into the OpenTelemetry ecosystem. Young described Alloy as an “observability Swiss Army knife” that gives you one tool to collect all your telemetry.

Öjeling also demonstrated Alloy’s live debugging capabilities, native Prometheus pipelines, and Fleet Management features. Fleet Management, she said, provides the convenience of having one place to get an overview of all of your collectors and configure them remotely. You can also use it to activate and deactivate pipelines without having to redeploy any collectors.

“This gives you control over the costs while at the same time gives you access to the data that you need when you need it,” she said.

Other references to how OpenTelemetry is impacting our internal development efforts were sprinkled throughout GrafanaCON 2025. For example, Staff Software Engineer Trevor Whitney mentioned how his team is using OpenTelemetry to improve Logs Drilldown, which is part of the Grafana Drilldown suite of apps that provide queryless, point-and-click experience so you can quickly find insights in your observability data.

“Grafana, as you all know, is a ‘big tent’ company. And so we’re really excited about this portability that OpenTelemetry is bringing to the observability community and with this new standard, which is a great way to add structure to your logs data and make your Drilldown come alive,” he said.

Prometheus + OTel: How Prometheus 3.0 addresses OTel compatibility

You might be asking yourself, Why would a Prometheus session get included in a blog post about OpenTelemetry? Well, it turns out there’s a lot of overlap in the two communities, with 70% of all organizations using Prometheus and OpenTelemetry in some capacity, and 34% using both in production environments, according to our third annual Observability Survey.

And getting those two projects to work together was a major focus of the Prometheus 3.0 release.

“OTel is gaining adoption,” said Product Manager Goutham Veeramachaneni. “We really like OTel, and if you go back just a year, using OTel with Prometheus was not great—well, it was painful. But over the past year we’ve had a huge push to support OTel natively in Prometheus.”

Veeramachaneni and Staff Software Engineer Carrie Edwards discussed a number of updates, including the addition of resource attributes, as well as support for OTLP, delta temporalities, and UTF-8. 

“Prometheus now has native support for OpenTelemetry, allowing users to push OTel metrics directly into Prometheus,” Veeramachaneni said. 

Experimenting with OTel: homelab fun with Beyla

Of course, it wouldn’t be GrafanaCON without some talk about fun side projects. And if you’re just looking to get started with OpenTelemetry, this homelab session was a good place to begin.

Veeramachaneni was joined by Principal Software Engineer Nikola Grcevski, and they demonstrated how Beyla can be used in homelabs, where applications are written in many languages and frameworks and you don’t have any control over their instrumentation. 

The pair showed how to deploy Beyla with Prometheus. They also discussed the “super power” that is OpenTelemetry’s Semantic Conventions, and they described how easy it is to use Beyla to observe your services. To check out the full session and see how you can start experimenting with OpenTelemetry at home, check out the full talk.

Grafana Labs’ OpenTelemetry contributions today and into the future

Grafana Labs works hard to maintain the OpenTelemetry language SDKs, Prometheus integrations, and OpenTelemetry Collector ecosystem as part of the OpenTelemetry community. We are a key part of the OpenTelemetry Profiling effort, as well as the new Browser and Mobile projects.

There’s also the aforementioned Beyla donation. We’re excited to see how integrating OpenTelemetry eBPF Instrumentation with profiling and other parts of OpenTelemetry can get us close to comprehensive observability that is completely automatic.

Pushing forward, we see a lot of value in the OpenTelemetry Semantic Conventions. Semantic Conventions form a contract between the telemetry and the concepts that the telemetry represents. Using weaver and other tools to generate better instrumentation APIs, automate dashboard creation, and improve AI coding agents, we think that a variety of projects coming out of the Semantic Convention Tooling SIG will be the surprise hit of OpenTelemetry in this coming year.

Grafana Cloud is the easiest way to get started with metrics, logs, traces, dashboards, and more. We have a generous forever-free tier and plans for every use case. Sign up for free now!