OBSERVABILITY / Grafana Labs
Grafana Pyroscope
Select SDK or Alloy profiling and inspect service-specific CPU, allocation and contention profiles.
“Continuously profile applications with Grafana Pyroscope and read the result as flame graphs. Covers three instrumentation paths — language SDK push (Go / Java / Python / Ruby / Node / .NET / Rust), Alloy eBPF auto-instrumentation (no code change, requires kernel 5.8+ with BTF), and SDK → Alloy receiver — plus ProfileQL queries, profile types (CPU / memory / allocations / goroutines / mutex), Grafana Cloud Profiles endpoint, and Span Profiles trace-to-profile linking. …”
01 / THE REASONING
Why this made the selection.
- It instruments an application for Grafana Pyroscope, either against a local server on port 4040 or Grafana Cloud with a numeric user and API key.
- A ready check and a labels query are the local proof that profiles arrived; Cloud still needs the stack id.
- The definition distinguishes three instrumentation paths, profile-type selectors and receiver versus application connectivity checks.
- Investigate CPU or allocation hotspots with the appropriate profile type
02 / THE REVIEW RECORD
What we actually inspected.
Source review has boundaries.
A clear record is more useful than a “safe” badge.
Material inspected
- skills/grafana-lgtm/pyroscope/SKILL.md
- README.md
- LICENSE
Our findings
- The Python example pushes to http://pyroscope:4040 and then checks /ready and /api/v1/labels.
- Grafana Cloud needs a numeric stack id and an API key, which the skill does not embed.
- Cluster-wide profiling is eBPF through Alloy, with a reload against http://localhost:12345.
- The skill does not query ProfileQL results as a substitute for the service appearing in Grafana.
- The definition distinguishes three instrumentation paths, profile-type selectors and receiver versus application connectivity checks.
- The troubleshooting section distinguishes an absent profile from a profile without frames, with separate Java format and Python-on-Alpine compatibility checks.
Not established by this review
- The resource was not executed and its tools were not called.
- No skill execution, authenticated service operation, or outcome benchmark was performed.
- Not executed: it was not installed or run.
The review applies to the material and revision named here. A newer upstream release can change its behavior.
03 / PUT IT TO WORK
Add a skill to your workflow.
DOCUMENTED COMMAND
npx skills add grafana/skills --skill pyroscope Run in your terminal to choose your AI tools, project or global scope, and installation method. This short command follows the current upstream version; the review covers the revision shown above.
- The README documents npx skills add grafana/skills --skill <name>. The table names this skill pyroscope.
- That command does not pin revision 1ccacf29049fde66637fe01ab93a83a772da323b. Inspect the installed SKILL.md against that revision before use.
- Do not apply the eBPF example on a host you are not authorised to profile.
Before you start
- Node.js and npm for the documented Skills installer
- A Pyroscope server or Grafana Cloud Profiles endpoint only if you run the instrumentation examples
04 / FOLLOW THE EVIDENCE
The source trail.
Our notes are separate from the original resource.
Check upstream before adopting a new version.
- Pinned skill definition ↗Checked
https://github.com/grafana/skills/blob/1ccacf29049fde66637fe01ab93a83a772da323b/skills/grafana-lgtm/pyroscope/SKILL.md
Supports: summary, whySelected, bestFor, limitations, install, access, review
https://github.com/grafana/skills/blob/1ccacf29049fde66637fe01ab93a83a772da323b/README.md
Supports: install
- Apache-2.0 licence ↗Checked
https://github.com/grafana/skills/blob/1ccacf29049fde66637fe01ab93a83a772da323b/LICENSE
Supports: license
https://raw.githubusercontent.com/grafana/skills/1ccacf29049fde66637fe01ab93a83a772da323b/skills/grafana-lgtm/pyroscope/SKILL.md
Supports: review.findings, bestFor, upstreamDescription