[ty] Detect flaky projects during ecosystem CI jobs#23178
Merged
Conversation
dcreager
added
internal
|
dcreager
force-pushed
the
dcreager/flake-detection
branch
from
4b99b1c to
47cd57e
Compare
dcreager
force-pushed
the
dcreager/flake-detection
branch
from
47cd57e to
cb7fe67
Compare
Typing conformance resultsNo changes detected ✅ |
|
dcreager
force-pushed
the
dcreager/flake-detection
branch
from
524fd18 to
23fb3be
Compare
dcreager
commented
Feb 9, 2026
Comment on lines
+81
to
+86
| --flaky-runs 10 \ | ||
| diff \ | ||
| --profile=profiling \ | ||
| --projects-old ruff/projects_old.txt \ | ||
| --projects-new ruff/projects_new.txt \ | ||
| --projects-flaky ruff/projects_flaky.txt \ |
Member
Author
There was a problem hiding this comment.
This is the job that runs for every PR (that has the ecosystem-analyzer label). Human people are blocked waiting for it to run, so we want to keep its runtime as short as possible. So we only run flake detection for a fixed set of projects. The 10 runs was chosen empirically — that's the number that gave stable results for the flaky projects mentioned below.
| ecosystem-analyzer \ | ||
| --verbose \ | ||
| --repository ruff \ | ||
| --flaky-runs 20 \ |
Member
Author
There was a problem hiding this comment.
This is the weekly CI job that runs in the background and does not block anyone's progress on any given PR. So I'm having it run more EXTREME flake detection against all projects. We can periodically check the results at https://ty-ecosystem-ext.pages.dev/ to see if we need to update the list.
dcreager
requested review from
AlexWaygood,
carljm and
sharkdp
as code owners
carljm
added a commit
to Hugo-Polloli/ruff
that referenced
this pull request
Feb 9, 2026
* main: [ty] Detect flaky projects during ecosystem CI jobs (astral-sh#23178) [ty] Disallow TypeVars within ClassVars (astral-sh#23184)
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
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.
This brings in an
ecosystem-analyzerupdate that detects when we produce flaky results for a project.We do this by running
tymultiple times against each project, looking for when we produce a different diagnostic at the same file/line/column location. If we do, we consider that diagnostic as flaky. ThediffandanalyzeHTML reports have been updated to show flakiness information. (A diagnostic changing from "flaky" to "not flaky", or vice versa, is now considered a change, as is a different set of diagnostic codes/messages appearing for a flaky diagnostic.)Since most projects aren't flaky 🤞, in the per-PR CI job, we only run flake detection against the projects that we known are flaky. For the others, we still run
tyonce and don't consider any of the diagnostics flaky. In the weekly background job, we run flake detection more aggressively, and on all projects, so that we can update the flaky project list when needed.