[flamegraph-basic] Flame Graph for Performance Profiling

[MarkusNeusinger]

The standard visualization for CPU/performance profiling, invented by Brendan Gregg. Stacked horizontal bars representing call stacks, width proportional to time spent. Every software developer knows this plot. Used in profiling tools across all major languages.

[claude]

✅ Spec ID: flamegraph-basic

Rationale: A flame graph is a well-established, widely recognized visualization for CPU/performance profiling data — no similar spec exists in the repository.

This is a valid and highly useful plot type invented by Brendan Gregg, standard across profiling tools in every major language ecosystem.

[github-actions]

✅ Specification Ready: flamegraph-basic

PR: #4678

---

specification.md

flamegraph-basic: Flame Graph for Performance Profiling

Description

A flame graph visualizes hierarchical call stack data from performance profiling, where each horizontal bar represents a function in the call stack and its width is proportional to the time (or samples) spent in that function. Stacks are layered bottom-to-top showing caller-to-callee relationships. Invented by Brendan Gregg, flame graphs are the standard visualization for identifying CPU bottlenecks and hot code paths across all major programming languages and profiling tools.

Applications

• Analyzing CPU profiling data to identify performance bottlenecks in application code
• Visualizing memory allocation call stacks to find sources of excessive allocation
• Comparing before/after profiling snapshots during performance optimization work

Data

• function (string) - Name of the function in the call stack
• stack (string) - Semicolon-delimited call stack path from root to leaf (e.g., main;process;compute)
• value (numeric) - Number of samples or time units spent in this stack frame
• Size: 50-500 unique stack traces
• Example: Simulated CPU profiling data with nested function call hierarchies and sample counts

Notes

• Each row of stacked bars represents a depth level in the call stack, with the root at the bottom
• Bar width encodes the proportion of total samples, not execution order (x-axis ordering is alphabetical or arbitrary, not temporal)
• Use a warm color palette (yellows, oranges, reds) following the conventional flame graph aesthetic
• Bars should be directly adjacent with minimal or no gaps between siblings at the same stack depth
• Include function name labels inside bars when the bar is wide enough to fit the text

---

Next: Add approved label to merge this specification to main.

---

🤖 spec-create workflow

[github-actions]

🚀 Specification Merged!

Specification: flamegraph-basic
PR: #4678 (merged)

The specification is now in main and synced to PostgreSQL.

---

Next: Add generate:{library} labels to start generating implementations.

Available libraries: matplotlib, seaborn, plotly, bokeh, altair, plotnine, pygal, highcharts, letsplot

Or use workflow dispatch for bulk operations.

---

🤖 spec-create workflow