gh-138122: Add thread status statistics to flamegraph profiler #141900
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
pablogsal
force-pushed
the
flamegraph-better
branch
from
c226154 to
3e0224f
Compare
pablogsal
force-pushed
the
flamegraph-better
branch
from
3e0224f to
21bbd24
Compare
…raph profiler The flamegraph profiler now collects thread state information (GIL held, GIL released, waiting for GIL, and garbage collection activity) during sampling and displays it in an interactive statistics bar. This helps users identify GIL contention and understand thread behavior patterns. Statistics are calculated both in aggregate and per-thread. When filtering the flamegraph to a specific thread, the display updates to show that thread's metrics. In GIL-only profiling mode, GIL-related statistics are hidden since they aren't meaningful in that context. The collection logic was refactored to process each sample in a single pass rather than iterating through threads multiple times. GC frame detection now uses a shared helper that consistently handles both tuple and object frame formats. Per-thread GC percentages now use total samples as the denominator instead of per-thread sample counts. This makes them directly comparable with aggregate statistics and easier to interpret when threads appear in different numbers of samples. The JavaScript includes null checks for DOM access and validates data availability before rendering. Tests verify the JSON structure, percentage calculations, and edge cases like zero samples, using only public APIs.
pablogsal
force-pushed
the
flamegraph-better
branch
from
21bbd24 to
9694d5a
Compare
Member
Author
… flamegraph profiler
Contributor
ivonastojanovic
left a comment
ivonastojanovic
left a comment
There was a problem hiding this comment.
Looks good to me! One small thing: for the percentages in the thread-stats-bar, maybe we should use the same rounding as the flamegraph nodes so everything is consistent.
Co-authored-by: ivonastojanovic <80911834+ivonastojanovic@users.noreply.github.com>
StanFromIreland
pushed a commit
to StanFromIreland/cpython
that referenced
this pull request
Dec 6, 2025
…ython#141900) Co-authored-by: ivonastojanovic <80911834+ivonastojanovic@users.noreply.github.com>
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
The flamegraph profiler now collects thread state information (GIL held, GIL released, waiting for GIL, and garbage collection activity) during sampling and displays it in an interactive statistics bar. This helps users identify GIL contention and understand thread behavior patterns.
Statistics are calculated both in aggregate and per-thread. When filtering the flamegraph to a specific thread, the display updates to show that thread's metrics. In GIL-only profiling mode, GIL-related statistics are hidden since they aren't meaningful in that context.
Per-thread GC percentages now use total samples as the denominator instead of per-thread sample counts. This makes them directly comparable with aggregate statistics and easier to interpret when threads appear in different numbers of samples.