Elastic’s donation of Universal Profiling agent to OpenTelemetry further solidifies profiling as core telemetry signal

Elastic has introduced that it could be donating its Common Profiling agent to the OpenTelemetry undertaking, setting the stage for profiling to turn into a fourth core telemetry sign along with logs, metrics, and tracing.

This follows OpenTelemetry’s announcement in March that it could be supporting profiling and was working in the direction of having a secure spec and implementation someday this 12 months.

Elastic’s agent profiles each line of code working on an organization’s machines, together with utility code, kernels, and third-party libraries. It’s all the time working within the background and may acquire knowledge about an utility over time.

It measures code effectivity throughout three classes: CPU utilization, CO2, and cloud value. In line with Elastic, this helps firms establish areas the place waste may be diminished or eradicated in order that they’ll optimize their programs.

Common Profiling at the moment helps numerous runtimes and languages, together with C/C++, Rust, Zig, Go, Java, Python, Ruby, PHP, Node.js, V8, Perl, and .NET.

“This contribution not solely boosts the standardization of steady profiling for observability but in addition accelerates the sensible adoption of profiling because the fourth key sign in OTel. Prospects get a vendor-agnostic manner of amassing profiling knowledge and enabling correlation with current indicators, like tracing, metrics, and logs, opening new potential for observability insights and a extra environment friendly troubleshooting expertise,” Elastic wrote in a weblog submit.

OpenTelemetry echoed these sentiments, saying: “This marks a big milestone in establishing profiling as a core telemetry sign in OpenTelemetry. Elastic’s eBPF primarily based profiling agent observes code throughout completely different programming languages and runtimes, third-party libraries, kernel operations, and system assets with low CPU and reminiscence overhead in manufacturing. Each, SREs and builders can now profit from these capabilities: rapidly figuring out efficiency bottlenecks, maximizing useful resource utilization, decreasing carbon footprint, and optimizing cloud spend.”