Hi Майя Гусарина,
When a service slows down, the trace shows where the time went, but not why. Is it a hot loop, a slow serializer, or a memory leak in a dependency? Finding out usually means reproducing the problem with a profiler attached, long after the incident is over.
With Dash0 Profiling, now generally available, you no longer have to.
It runs continuously through an eBPF agent on each node, with no code changes. It supports all major programming languages across Kubernetes, Linux hosts, VMs and bare metal.
The profile from the minute something went wrong is already there, next to the traces, logs and metrics from the same pod.
- Built for agents. Agent0 and coding agents like Claude Code read the same profiles through Dash0's MCP server, so a suggested fix starts from measured cost, not a guess.
- Open by design. Built by the team behind Polar Signals, who contribute to the OpenTelemetry Profiling Working Group. Parca remains open source, and Dash0 accepts any profile type in the OpenTelemetry format.
Try it on your own services with the 14-day free trial.