Repository navigation
ci: rebalance test shards 4→6 to cut wall-clock - #279
Merged
Merged
Conversation
The 45 test projects were split across 4 shards; shard 0 ran 24m while shard 1 ran 10m — everyone waited on the outlier. Raising SHARD_TOTAL 4→6 spreads the ~38 light projects over 6 shards instead of 4 (the extra light projects that padded shard 0 move off it), while the heavy/light dual-counter still keeps the 3 ALC-heaviest (AI/FutuRe/Orleans) on distinct shards and only one shard doubles up (Acme+Orleans). Wall-clock ≈ build (~10m) + slowest shard, so cutting the long pole cuts the run. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Contributor
There was a problem hiding this comment.
Pull request overview
This PR reduces CI wall-clock time for the .NET test workflow by increasing the number of parallel test shards, improving distribution of light test projects while preserving the existing heavy/light weighted sharding strategy.
Changes:
- Increased the test job matrix from 4 shards to 6 shards (
shard: [0..5]). - Updated
SHARD_TOTALfrom4to6so the bash sharding logic assigns projects across 6 runners. - Updated inline workflow comments/diagram to reflect the new shard count and rationale.
💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.
Contributor
Test Results (shard 2)696 tests - 416 694 ✅ - 417 1m 48s ⏱️ - 4m 48s Results for commit 3c02d31. ± Comparison against base commit c3d15ed. This pull request removes 463 and adds 47 tests. Note that renamed tests count towards both. |
Contributor
Test Results (shard 3)788 tests - 951 788 ✅ - 950 3m 25s ⏱️ - 3m 31s Results for commit 3c02d31. ± Comparison against base commit c3d15ed. This pull request removes 1067 and adds 116 tests. Note that renamed tests count towards both. |
Contributor
Test Results (shard 1)1 029 tests +88 1 028 ✅ +267 3m 29s ⏱️ - 1m 1s Results for commit 3c02d31. ± Comparison against base commit c3d15ed. This pull request removes 843 and adds 931 tests. Note that renamed tests count towards both.This pull request removes 180 skipped tests and adds 1 skipped test. Note that renamed tests count towards both. |
Contributor
Test Results (shard 4) 10 files 10 suites 4m 30s ⏱️ Results for commit 3c02d31. |
Contributor
Test Results (shard 5)862 tests 682 ✅ 4m 14s ⏱️ Results for commit 3c02d31. |
Contributor
Test Results (shard 0) 10 files - 5 10 suites - 5 4m 56s ⏱️ +38s Results for commit 3c02d31. ± Comparison against base commit c3d15ed. This pull request removes 1051 and adds 138 tests. Note that renamed tests count towards both.♻️ This comment has been updated with latest results. |
Contributor
Test Results 56 files ±0 56 suites ±0 22m 25s ⏱️ +3s Results for commit 3c02d31. ± Comparison against base commit c3d15ed. This pull request removes 9 tests.♻️ This comment has been updated with latest results. |
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
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.
What
Raises
SHARD_TOTAL4 → 6 indotnet-test.yml(matrix + env).Why
From a real run: the 45 test projects split across 4 shards left shard 0 at 24m while shard 1 ran 10m — the whole run waits on the outlier (wall-clock ≈ build ~10m + slowest shard ≈ 34m). Spreading the ~38 light projects over 6 shards instead of 4 moves the extra light projects off the long-pole shard.
The existing heavy/light dual-counter balancing is preserved: at 6 shards the 7 heavies still spread ≤2/shard, the 3 ALC-heaviest (AI/FutuRe/Orleans) still land on distinct shards, and only one shard doubles up (Acme+Orleans). Tests stay serial within a shard (
maxParallelThreads:1— required for correctness), so more shards is the right lever.The collector job fans in via
testResults-shard*glob +needs: test, so it adapts to 6 shards with no change.🤖 Generated with Claude Code
EOF