Grafana Cloud Profiles Aims to Unify Observability for CloudOps

Observability vendor Grafana Labs announced the availability of a new service on Aug. 9 that could help organizations get a much better view of how applications are running.

The Grafana Cloud Profiles service is now generally available, bringing new observability capabilities to enterprises. The new service is based on technology from Pyroscope, a vendor that Grafana acquired in March. Pyroscope has an open source project, simply known as Pyroscope, that is a continuous profiling platform for applications. With Cloud Profiles, Grafana claims that users can more easily find the root causes of different types of incidents and inefficiencies with profiles and metrics.

"Grafana Cloud Profiles introduces a shift in how we understand resource utilization and performance of systems often containing millions of lines of code," Ryan Perry, director of engineering at Grafana Labs and former CEO of Pyroscope, told ITPro Today.

Traditionally, Perry explained, metrics, logs, and traces provide valuable insights into system health and performance, but not how that performance is attributed to the underlying lines of code. The continuous profiling approach bridges that gap. At a system-wide level, profiling gives users a unique view of code performance/patterns and highlights hot spots for resource consumption, he added. 

Related:Observability vs. Monitoring: Who Needs Which When?

Perry-Grafana

At the same time, if you need to dig deeper, Grafana's profiling database is equipped to dive into the performance of individual services and at times even individual log lines, metrics, or trace spans, Perry said. Without profiling, achieving this detailed perspective would be almost impossible.

Grafana Cloud Profiles Goes Beyond Just Pyroscope for Continuous Profiling

Grafana Cloud Profiles isn't just the open source Pyroscope project running in the Grafana Labs cloud platform.

"Rather than simply embedding our profiling app into Grafana, we've enhanced its capability through weaving it into the product alongside other signals," Perry said.

For example, users can now create dashboards that show not just that a server is running out of memory, but the profile that points to where the memory leak is located in code, he said. In the Explore tab, now users don't just see a list of slow traces, but also the profiles associated with them that point to exactly why they're slow.

"With Grafana Cloud Profiles, profiling is much more than just another signal; it enhances the value of the entire suite of observability tools," Perry said.

The Challenges of Cloud Observability

Organizations face a number of challenges regarding cloud observability.

Among the biggest issues, according to Perry, is the exponential growth in system complexity. He noted that often, traditional tools view logs, metrics, and traces as separate, siloed tools. This leads many into the trap of getting a tool and then forcing it to answer questions about their systems, the common problem of "when you have a hammer everything looks like a nail." Instead of this tool-first approach, Perry argued that a more holistic view is much more effective.

"It's not just about using logs, metrics, traces, profiles individually, but understanding how much more you can learn about your systems when you use them collectively together," he said.