grafana/pyroscope

Pyroscope 2.0 makes the new v2 architecture the default. Profiles are written directly to object storage, removing the need for in-memory ingesters and local disks - simplifying operations and lowering resource usage at scale. Existing v1 deployments can opt in via a flag and migrate without data loss.

Read the 2.0 release notes and the v2 architecture overview. Upgrading from v1? See the migration guide.

Want the full story? Watch the GrafanaCON 2026 talk Pyroscope 2.0: Continuous Profiling Architecture Deep Dive.

Grafana Pyroscope is a continuous profiling platform designed to surface performance insights from your applications, helping you optimize resource usage such as CPU, memory, and I/O operations. With Pyroscope, you can both proactively and reactively address performance bottlenecks across your system.

The typical use cases are:

• Proactive: Reducing resource consumption, improving application performance, or preventing latency issues.
• Reactive: Quickly resolving incidents with line-level detail and debugging active CPU, memory, or I/O bottlenecks.

Pyroscope provides powerful tools to give you a comprehensive view of your application's behavior while allowing you to drill down into specific services for more targeted root cause analysis.

Pyroscope consists of three main components:

• Pyroscope Server: Stores and processes profiling data and serves queries.
• Clients and instrumentation: Profiling data reaches the server in several ways: the Pyroscope SDKs (push), Grafana Alloy (pull or push), or OTLP from OpenTelemetry-compatible sources such as the OpenTelemetry eBPF profiler.
• Grafana Profiles Drilldown: A queryless, intuitive UI for visualizing and analyzing profiling data (formerly Explore Profiles).

Under the hood, Pyroscope v2 writes profiles straight to object storage—no ingesters, no local disk. The animations below trace the three parts of the architecture. For the details behind each component, see the v2 architecture documentation.

Write path — profiles are routed by service and written straight to object storage:

Compaction — compaction-workers merge small segments into larger blocks in the background:

Read path — queries fan out across object storage to build flame graphs in Grafana Profiles Drilldown:

docker run -it -p 4040:4040 grafana/pyroscope

brew install pyroscope-io/brew/pyroscope
brew services start pyroscope

Download the archive for your operating system and architecture from the latest release, unpack it, and run the binary:

tar xvf pyroscope_*.tar.gz
./pyroscope

Pyroscope listens on port 4040. For Kubernetes/Helm, Linux packages, building from source, and full configuration options, see the Get started guide and the server documentation.

Grafana Profiles Drilldown (formerly Explore Profiles) is the primary, queryless way to visualize and analyze your profiling data.

Profiles Drilldown is pre-installed and is the default way to explore your profiles – all you need to do is start sending data.

For more information on how to use Pyroscope with other programming languages, install it on Linux, or use it in a production environment, check out our documentation:

• Getting Started
• Deployment Guide
• Pyroscope v2 Architecture

You can send profiles to Pyroscope with the language SDKs, with Grafana Alloy, or over OTLP from OpenTelemetry-compatible sources such as the OpenTelemetry eBPF profiler.

For more documentation on how to add the Pyroscope SDK to your code, see the client documentation on our website or find language-specific examples and documentation below:

Golang Documentation Examples	Java Documentation Examples	Python Documentation Examples	Ruby Documentation Examples
Node.js Documentation Examples	.NET Documentation Examples	eBPF Documentation Examples	Rust Documentation Examples

Our documentation contains the most recent list of supported languages and also an overview over what profiling types are supported per language.

Let us know what other integrations you want to see in our issues or in our slack.

Pyroscope is possible thanks to the excellent work of many people, including but not limited to:

• Brendan Gregg — inventor of Flame Graphs
• Julia Evans — creator of rbspy — sampling profiler for Ruby
• Vladimir Agafonkin — creator of flamebearer — fast flame graph renderer
• Ben Frederickson — creator of py-spy — sampling profiler for Python
• Adam Saponara — creator of phpspy — sampling profiler for PHP
• Alexei Starovoitov, Daniel Borkmann, and many others who made BPF based profiling in Linux kernel possible
• Jamie Wong — creator of speedscope — interactive flame graph visualizer

To start contributing, check out our Contributing Guide