Remove deploy steps from CI #1
Merged
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
Signed-off-by: Marco Pracucci <[email protected]>
56quarters
approved these changes
Jul 14, 2021
tomwilkie
approved these changes
Jul 14, 2021
18 tasks
radudd
added a commit
to radudd/mimir
that referenced
this pull request
Sep 9, 2022
Co-authored-by: Dimitar Dimitrov <[email protected]>
gotjosh
added a commit
that referenced
this pull request
Jun 25, 2024
This PR introduces two subtle but important changes (it's mostly a proposal and feel free to argue on this): 1. We now wait until we stop _all_ of the managers. This is important, because managers don't stop rule evaluation that are in-flight, so we'll wait until we finish any sort of rule group evaluation before considering the "tenant removed". 2. Also, once we're done with rule evaluation we _actually_ stop the notifiers. Notifiers weren't stopped before which was mostly a good thing (as we would always make sure that any alerts were getting flushed). All in all, #1 is mostly up for discussion. However, I think #2 should go in as it makes the behaviour explicit. Signed-off-by: gotjosh <[email protected]>
4 tasks
gotjosh
added a commit
that referenced
this pull request
Jun 28, 2024
This PR introduces two subtle but important changes (it's mostly a proposal and feel free to argue on this): 1. We now wait until we stop _all_ of the managers. This is important, because managers don't stop rule evaluation that are in-flight, so we'll wait until we finish any sort of rule group evaluation before considering the "tenant removed". 2. Also, once we're done with rule evaluation we _actually_ stop the notifiers. Notifiers weren't stopped before which was mostly a good thing (as we would always make sure that any alerts were getting flushed). All in all, #1 is mostly up for discussion. However, I think #2 should go in as it makes the behaviour explicit. Signed-off-by: gotjosh <[email protected]>
This was referenced Jun 9, 2025
Closed
pracucci
added a commit
that referenced
this pull request
Jun 23, 2025
#### What this PR does This adds a bloom filter optimization on top of #11819 Benchmark compared to that branch: ``` $ benchstat -ignore .name shortcuts bloom-16-buckets goos: darwin goarch: arm64 pkg: github.com/grafana/mimir/pkg/compactor cpu: Apple M3 Pro │ shortcuts │ bloom-16-buckets │ │ sec/op │ sec/op vs base │ */Block-10/#00-12 26.85µ ± 6% 28.56µ ± 2% +6.36% (p=0.015 n=6) */Block-10/#1-12 81.48µ ± 4% 84.76µ ± 0% +4.03% (p=0.002 n=6) */Block-100/#00-12 365.6µ ± 2% 317.9µ ± 0% -13.05% (p=0.002 n=6) */Block-100/#1-12 1091.5µ ± 1% 958.5µ ± 5% -12.19% (p=0.002 n=6) */Block-1000/#00-12 15.122m ± 8% 6.804m ± 2% -55.01% (p=0.002 n=6) */Block-1000/#1-12 46.47m ± 4% 20.80m ± 5% -55.23% (p=0.002 n=6) */Block-10000/#00-12 2925.2m ± 5% 638.9m ± 1% -78.16% (p=0.002 n=6) */Block-10000/#1-12 8.722 ± 2% 2.576 ± 2% -70.47% (p=0.002 n=6) geomean 7.931m 4.512m -43.11% ``` Benchmark compared to main: ``` $ benchstat -ignore .name main shortcuts bloom goos: darwin goarch: arm64 pkg: github.com/grafana/mimir/pkg/compactor cpu: Apple M3 Pro │ main │ shortcuts │ bloom │ │ sec/op │ sec/op vs base │ sec/op vs base │ */Block-10/#00-12 17.30µ ± 2% 26.85µ ± 6% +55.22% (p=0.002 n=6) 30.99µ ± 3% +79.14% (p=0.002 n=6) */Block-10/#1-12 50.17µ ± 1% 81.48µ ± 4% +62.41% (p=0.002 n=6) 90.49µ ± 2% +80.36% (p=0.002 n=6) */Block-100/#00-12 378.8µ ± 0% 365.6µ ± 2% -3.47% (p=0.002 n=6) 360.7µ ± 2% -4.77% (p=0.002 n=6) */Block-100/#1-12 1.133m ± 1% 1.092m ± 1% -3.67% (p=0.002 n=6) 1.086m ± 1% -4.20% (p=0.002 n=6) */Block-1000/#00-12 28.244m ± 1% 15.122m ± 8% -46.46% (p=0.002 n=6) 7.974m ± 2% -71.77% (p=0.002 n=6) */Block-1000/#1-12 90.46m ± 13% 46.47m ± 4% -48.64% (p=0.002 n=6) 24.95m ± 9% -72.42% (p=0.002 n=6) */Block-10000/#00-12 7.936 ± 7% 2.925 ± 5% -63.14% (p=0.002 n=6) 1.393 ± 3% -82.44% (p=0.002 n=6) */Block-10000/#1-12 24.905 ± 9% 8.722 ± 2% -64.98% (p=0.002 n=6) 4.256 ± 3% -82.91% (p=0.002 n=6) geomean 10.82m 7.931m -26.71% 5.808m -46.33% ``` <details> <summary>Benchmark results with different buckets counts</summary> I first tried with 256 buckets, and that was slower than 16: ``` $ benchstat -ignore .name shortcuts bloom-16-buckets bloom-256-buckets goos: darwin goarch: arm64 pkg: github.com/grafana/mimir/pkg/compactor cpu: Apple M3 Pro │ shortcuts │ bloom-16-buckets │ bloom-256-buckets │ │ sec/op │ sec/op vs base │ sec/op vs base │ */Block-10/#00-12 26.85µ ± 6% 28.56µ ± 2% +6.36% (p=0.015 n=6) 30.99µ ± 3% +15.41% (p=0.002 n=6) */Block-10/#1-12 81.48µ ± 4% 84.76µ ± 0% +4.03% (p=0.002 n=6) 90.49µ ± 2% +11.05% (p=0.002 n=6) */Block-100/#00-12 365.6µ ± 2% 317.9µ ± 0% -13.05% (p=0.002 n=6) 360.7µ ± 2% -1.35% (p=0.015 n=6) */Block-100/#1-12 1091.5µ ± 1% 958.5µ ± 5% -12.19% (p=0.002 n=6) 1085.5µ ± 1% ~ (p=0.180 n=6) */Block-1000/#00-12 15.122m ± 8% 6.804m ± 2% -55.01% (p=0.002 n=6) 7.974m ± 2% -47.27% (p=0.002 n=6) */Block-1000/#1-12 46.47m ± 4% 20.80m ± 5% -55.23% (p=0.002 n=6) 24.95m ± 9% -46.31% (p=0.002 n=6) */Block-10000/#00-12 2925.2m ± 5% 638.9m ± 1% -78.16% (p=0.002 n=6) 1393.5m ± 3% -52.36% (p=0.002 n=6) */Block-10000/#1-12 8.722 ± 2% 2.576 ± 2% -70.47% (p=0.002 n=6) 4.256 ± 3% -51.21% (p=0.002 n=6) geomean 7.931m 4.512m -43.11% 5.808m -26.76% ``` I also tried with a single uint, which gave me suprisingly similar number: ``` $ benchstat -ignore .name shortcuts bloom-16-buckets bloom-single goos: darwin goarch: arm64 pkg: github.com/grafana/mimir/pkg/compactor cpu: Apple M3 Pro │ shortcuts │ bloom-16-buckets │ bloom-single │ │ sec/op │ sec/op vs base │ sec/op vs base │ */Block-10/#00-12 26.85µ ± 6% 28.56µ ± 2% +6.36% (p=0.015 n=6) 27.50µ ± 1% ~ (p=0.180 n=6) */Block-10/#1-12 81.48µ ± 4% 84.76µ ± 0% +4.03% (p=0.002 n=6) 82.25µ ± 4% ~ (p=0.394 n=6) */Block-100/#00-12 365.6µ ± 2% 317.9µ ± 0% -13.05% (p=0.002 n=6) 313.4µ ± 4% -14.28% (p=0.002 n=6) */Block-100/#1-12 1091.5µ ± 1% 958.5µ ± 5% -12.19% (p=0.002 n=6) 947.3µ ± 6% -13.22% (p=0.002 n=6) */Block-1000/#00-12 15.122m ± 8% 6.804m ± 2% -55.01% (p=0.002 n=6) 6.712m ± 2% -55.61% (p=0.002 n=6) */Block-1000/#1-12 46.47m ± 4% 20.80m ± 5% -55.23% (p=0.002 n=6) 20.34m ± 2% -56.23% (p=0.002 n=6) */Block-10000/#00-12 2925.2m ± 5% 638.9m ± 1% -78.16% (p=0.002 n=6) 621.9m ± 0% -78.74% (p=0.002 n=6) */Block-10000/#1-12 8.722 ± 2% 2.576 ± 2% -70.47% (p=0.002 n=6) 2.504 ± 2% -71.29% (p=0.002 n=6) geomean 7.931m 4.512m -43.11% 4.409m -44.41% ``` It seems that using 4 or 8 buckets doesn't speed up the result, but 32 slows it down, so I've chosen 16 buckets as the magic number for our fake microbenchmark 🤷 ``` $ benchstat -ignore .name bloom-16-buckets bloom-4-buckets bloom-8-buckets bloom-32-buckets goos: darwin goarch: arm64 pkg: github.com/grafana/mimir/pkg/compactor cpu: Apple M3 Pro │ bloom-16-buckets │ bloom-4-buckets │ bloom-8-buckets │ bloom-32-buckets │ │ sec/op │ sec/op vs base │ sec/op vs base │ sec/op vs base │ */Block-10/#00-12 28.56µ ± 2% 28.38µ ± 7% ~ (p=0.699 n=6) 28.54µ ± 1% ~ (p=0.937 n=6) 28.47µ ± 2% ~ (p=0.240 n=6) */Block-10/#1-12 84.76µ ± 0% 84.02µ ± 4% ~ (p=0.132 n=6) 85.43µ ± 3% +0.78% (p=0.041 n=6) 84.68µ ± 1% ~ (p=0.937 n=6) */Block-100/#00-12 317.9µ ± 0% 321.5µ ± 3% ~ (p=0.589 n=6) 319.2µ ± 4% ~ (p=0.132 n=6) 319.1µ ± 4% +0.37% (p=0.041 n=6) */Block-100/#1-12 958.5µ ± 5% 953.7µ ± 7% ~ (p=0.394 n=6) 955.2µ ± 2% ~ (p=0.310 n=6) 959.1µ ± 1% ~ (p=0.818 n=6) */Block-1000/#00-12 6.804m ± 2% 6.721m ± 1% -1.22% (p=0.002 n=6) 6.762m ± 1% -0.61% (p=0.026 n=6) 7.022m ± 4% +3.21% (p=0.002 n=6) */Block-1000/#1-12 20.80m ± 5% 20.54m ± 0% -1.25% (p=0.002 n=6) 20.70m ± 1% ~ (p=0.093 n=6) 21.72m ± 2% +4.43% (p=0.041 n=6) */Block-10000/#00-12 638.9m ± 1% 633.8m ± 2% ~ (p=0.093 n=6) 631.3m ± 1% -1.18% (p=0.009 n=6) 703.4m ± 44% +10.09% (p=0.002 n=6) */Block-10000/#1-12 2.576 ± 2% 2.543 ± 2% ~ (p=0.180 n=6) 2.562 ± 2% ~ (p=0.065 n=6) 2.832 ± 23% +9.95% (p=0.002 n=6) geomean 4.512m 4.481m -0.68% 4.501m -0.26% 4.665m +3.38% ``` </details> #### Which issue(s) this PR fixes or relates to Ref: #11819 --------- Signed-off-by: Oleg Zaytsev <[email protected]> Signed-off-by: Marco Pracucci <[email protected]> Co-authored-by: Marco Pracucci <[email protected]>
pracucci
added a commit
that referenced
this pull request
Jun 23, 2025
#### What this PR does This adds a bloom filter optimization on top of #11819 Benchmark compared to that branch: ``` $ benchstat -ignore .name shortcuts bloom-16-buckets goos: darwin goarch: arm64 pkg: github.com/grafana/mimir/pkg/compactor cpu: Apple M3 Pro │ shortcuts │ bloom-16-buckets │ │ sec/op │ sec/op vs base │ */Block-10/#00-12 26.85µ ± 6% 28.56µ ± 2% +6.36% (p=0.015 n=6) */Block-10/#1-12 81.48µ ± 4% 84.76µ ± 0% +4.03% (p=0.002 n=6) */Block-100/#00-12 365.6µ ± 2% 317.9µ ± 0% -13.05% (p=0.002 n=6) */Block-100/#1-12 1091.5µ ± 1% 958.5µ ± 5% -12.19% (p=0.002 n=6) */Block-1000/#00-12 15.122m ± 8% 6.804m ± 2% -55.01% (p=0.002 n=6) */Block-1000/#1-12 46.47m ± 4% 20.80m ± 5% -55.23% (p=0.002 n=6) */Block-10000/#00-12 2925.2m ± 5% 638.9m ± 1% -78.16% (p=0.002 n=6) */Block-10000/#1-12 8.722 ± 2% 2.576 ± 2% -70.47% (p=0.002 n=6) geomean 7.931m 4.512m -43.11% ``` Benchmark compared to main: ``` $ benchstat -ignore .name main shortcuts bloom goos: darwin goarch: arm64 pkg: github.com/grafana/mimir/pkg/compactor cpu: Apple M3 Pro │ main │ shortcuts │ bloom │ │ sec/op │ sec/op vs base │ sec/op vs base │ */Block-10/#00-12 17.30µ ± 2% 26.85µ ± 6% +55.22% (p=0.002 n=6) 30.99µ ± 3% +79.14% (p=0.002 n=6) */Block-10/#1-12 50.17µ ± 1% 81.48µ ± 4% +62.41% (p=0.002 n=6) 90.49µ ± 2% +80.36% (p=0.002 n=6) */Block-100/#00-12 378.8µ ± 0% 365.6µ ± 2% -3.47% (p=0.002 n=6) 360.7µ ± 2% -4.77% (p=0.002 n=6) */Block-100/#1-12 1.133m ± 1% 1.092m ± 1% -3.67% (p=0.002 n=6) 1.086m ± 1% -4.20% (p=0.002 n=6) */Block-1000/#00-12 28.244m ± 1% 15.122m ± 8% -46.46% (p=0.002 n=6) 7.974m ± 2% -71.77% (p=0.002 n=6) */Block-1000/#1-12 90.46m ± 13% 46.47m ± 4% -48.64% (p=0.002 n=6) 24.95m ± 9% -72.42% (p=0.002 n=6) */Block-10000/#00-12 7.936 ± 7% 2.925 ± 5% -63.14% (p=0.002 n=6) 1.393 ± 3% -82.44% (p=0.002 n=6) */Block-10000/#1-12 24.905 ± 9% 8.722 ± 2% -64.98% (p=0.002 n=6) 4.256 ± 3% -82.91% (p=0.002 n=6) geomean 10.82m 7.931m -26.71% 5.808m -46.33% ``` <details> <summary>Benchmark results with different buckets counts</summary> I first tried with 256 buckets, and that was slower than 16: ``` $ benchstat -ignore .name shortcuts bloom-16-buckets bloom-256-buckets goos: darwin goarch: arm64 pkg: github.com/grafana/mimir/pkg/compactor cpu: Apple M3 Pro │ shortcuts │ bloom-16-buckets │ bloom-256-buckets │ │ sec/op │ sec/op vs base │ sec/op vs base │ */Block-10/#00-12 26.85µ ± 6% 28.56µ ± 2% +6.36% (p=0.015 n=6) 30.99µ ± 3% +15.41% (p=0.002 n=6) */Block-10/#1-12 81.48µ ± 4% 84.76µ ± 0% +4.03% (p=0.002 n=6) 90.49µ ± 2% +11.05% (p=0.002 n=6) */Block-100/#00-12 365.6µ ± 2% 317.9µ ± 0% -13.05% (p=0.002 n=6) 360.7µ ± 2% -1.35% (p=0.015 n=6) */Block-100/#1-12 1091.5µ ± 1% 958.5µ ± 5% -12.19% (p=0.002 n=6) 1085.5µ ± 1% ~ (p=0.180 n=6) */Block-1000/#00-12 15.122m ± 8% 6.804m ± 2% -55.01% (p=0.002 n=6) 7.974m ± 2% -47.27% (p=0.002 n=6) */Block-1000/#1-12 46.47m ± 4% 20.80m ± 5% -55.23% (p=0.002 n=6) 24.95m ± 9% -46.31% (p=0.002 n=6) */Block-10000/#00-12 2925.2m ± 5% 638.9m ± 1% -78.16% (p=0.002 n=6) 1393.5m ± 3% -52.36% (p=0.002 n=6) */Block-10000/#1-12 8.722 ± 2% 2.576 ± 2% -70.47% (p=0.002 n=6) 4.256 ± 3% -51.21% (p=0.002 n=6) geomean 7.931m 4.512m -43.11% 5.808m -26.76% ``` I also tried with a single uint, which gave me suprisingly similar number: ``` $ benchstat -ignore .name shortcuts bloom-16-buckets bloom-single goos: darwin goarch: arm64 pkg: github.com/grafana/mimir/pkg/compactor cpu: Apple M3 Pro │ shortcuts │ bloom-16-buckets │ bloom-single │ │ sec/op │ sec/op vs base │ sec/op vs base │ */Block-10/#00-12 26.85µ ± 6% 28.56µ ± 2% +6.36% (p=0.015 n=6) 27.50µ ± 1% ~ (p=0.180 n=6) */Block-10/#1-12 81.48µ ± 4% 84.76µ ± 0% +4.03% (p=0.002 n=6) 82.25µ ± 4% ~ (p=0.394 n=6) */Block-100/#00-12 365.6µ ± 2% 317.9µ ± 0% -13.05% (p=0.002 n=6) 313.4µ ± 4% -14.28% (p=0.002 n=6) */Block-100/#1-12 1091.5µ ± 1% 958.5µ ± 5% -12.19% (p=0.002 n=6) 947.3µ ± 6% -13.22% (p=0.002 n=6) */Block-1000/#00-12 15.122m ± 8% 6.804m ± 2% -55.01% (p=0.002 n=6) 6.712m ± 2% -55.61% (p=0.002 n=6) */Block-1000/#1-12 46.47m ± 4% 20.80m ± 5% -55.23% (p=0.002 n=6) 20.34m ± 2% -56.23% (p=0.002 n=6) */Block-10000/#00-12 2925.2m ± 5% 638.9m ± 1% -78.16% (p=0.002 n=6) 621.9m ± 0% -78.74% (p=0.002 n=6) */Block-10000/#1-12 8.722 ± 2% 2.576 ± 2% -70.47% (p=0.002 n=6) 2.504 ± 2% -71.29% (p=0.002 n=6) geomean 7.931m 4.512m -43.11% 4.409m -44.41% ``` It seems that using 4 or 8 buckets doesn't speed up the result, but 32 slows it down, so I've chosen 16 buckets as the magic number for our fake microbenchmark 🤷 ``` $ benchstat -ignore .name bloom-16-buckets bloom-4-buckets bloom-8-buckets bloom-32-buckets goos: darwin goarch: arm64 pkg: github.com/grafana/mimir/pkg/compactor cpu: Apple M3 Pro │ bloom-16-buckets │ bloom-4-buckets │ bloom-8-buckets │ bloom-32-buckets │ │ sec/op │ sec/op vs base │ sec/op vs base │ sec/op vs base │ */Block-10/#00-12 28.56µ ± 2% 28.38µ ± 7% ~ (p=0.699 n=6) 28.54µ ± 1% ~ (p=0.937 n=6) 28.47µ ± 2% ~ (p=0.240 n=6) */Block-10/#1-12 84.76µ ± 0% 84.02µ ± 4% ~ (p=0.132 n=6) 85.43µ ± 3% +0.78% (p=0.041 n=6) 84.68µ ± 1% ~ (p=0.937 n=6) */Block-100/#00-12 317.9µ ± 0% 321.5µ ± 3% ~ (p=0.589 n=6) 319.2µ ± 4% ~ (p=0.132 n=6) 319.1µ ± 4% +0.37% (p=0.041 n=6) */Block-100/#1-12 958.5µ ± 5% 953.7µ ± 7% ~ (p=0.394 n=6) 955.2µ ± 2% ~ (p=0.310 n=6) 959.1µ ± 1% ~ (p=0.818 n=6) */Block-1000/#00-12 6.804m ± 2% 6.721m ± 1% -1.22% (p=0.002 n=6) 6.762m ± 1% -0.61% (p=0.026 n=6) 7.022m ± 4% +3.21% (p=0.002 n=6) */Block-1000/#1-12 20.80m ± 5% 20.54m ± 0% -1.25% (p=0.002 n=6) 20.70m ± 1% ~ (p=0.093 n=6) 21.72m ± 2% +4.43% (p=0.041 n=6) */Block-10000/#00-12 638.9m ± 1% 633.8m ± 2% ~ (p=0.093 n=6) 631.3m ± 1% -1.18% (p=0.009 n=6) 703.4m ± 44% +10.09% (p=0.002 n=6) */Block-10000/#1-12 2.576 ± 2% 2.543 ± 2% ~ (p=0.180 n=6) 2.562 ± 2% ~ (p=0.065 n=6) 2.832 ± 23% +9.95% (p=0.002 n=6) geomean 4.512m 4.481m -0.68% 4.501m -0.26% 4.665m +3.38% ``` </details> #### Which issue(s) this PR fixes or relates to Ref: #11819 --------- Signed-off-by: Oleg Zaytsev <[email protected]> Signed-off-by: Marco Pracucci <[email protected]> Co-authored-by: Marco Pracucci <[email protected]>
pracucci
added a commit
that referenced
this pull request
Jun 26, 2025
#### What this PR does This PR has been be authored by @colega and myself. We had an issue where a tenant compaction was lagging behind, and their number blocks reached the monster number of 250K (from a baseline of 10K). When that happened, we found out that the compactor planning was extremely slow (3 hours), and that the slow down came from `ShardAwareDeduplicateFilter`. In this PR we introduce some small optimizations to `ShardAwareDeduplicateFilter` for tenants with a very large number of blocks. The impact on performance is not impressive, but we got a -33% reduction (from 3h30m to 2h15m) for the real world scenario. Existing benchmark result: ``` goos: darwin goarch: arm64 pkg: github.com/grafana/mimir/pkg/compactor cpu: Apple M3 Pro │ before.txt │ after-with-bloom-filter-and-without-list.txt │ │ sec/op │ sec/op vs base │ DeduplicateFilter_Filter/Block-10/#00-11 17.64µ ± 14% 25.23µ ± 3% +43.04% (p=0.002 n=6) DeduplicateFilter_Filter/Block-10/#1-11 53.33µ ± 1% 76.36µ ± 1% +43.18% (p=0.002 n=6) DeduplicateFilter_Filter/Block-100/#00-11 383.6µ ± 4% 294.2µ ± 1% -23.30% (p=0.002 n=6) DeduplicateFilter_Filter/Block-100/#1-11 1153.2µ ± 1% 882.1µ ± 1% -23.51% (p=0.002 n=6) DeduplicateFilter_Filter/Block-1000/#00-11 28.346m ± 1% 7.578m ± 6% -73.26% (p=0.002 n=6) DeduplicateFilter_Filter/Block-1000/#1-11 90.15m ± 1% 23.02m ± 1% -74.46% (p=0.002 n=6) DeduplicateFilter_Filter/Block-10000/#00-11 6.820 ± 5% 1.237 ± 1% -81.86% (p=0.002 n=6) DeduplicateFilter_Filter/Block-10000/#1-11 20.710 ± 2% 3.723 ± 1% -82.02% (p=0.002 n=6) geomean 10.52m 5.018m -52.31% │ before.txt │ after-with-bloom-filter-and-without-list.txt │ │ B/op │ B/op vs base │ DeduplicateFilter_Filter/Block-10/#00-11 24.87Ki ± 0% 30.59Ki ± 0% +23.00% (p=0.002 n=6) DeduplicateFilter_Filter/Block-10/#1-11 74.61Ki ± 0% 91.76Ki ± 0% +23.00% (p=0.002 n=6) DeduplicateFilter_Filter/Block-100/#00-11 247.4Ki ± 0% 299.5Ki ± 0% +21.07% (p=0.002 n=6) DeduplicateFilter_Filter/Block-100/#1-11 742.6Ki ± 0% 898.9Ki ± 0% +21.05% (p=0.002 n=6) DeduplicateFilter_Filter/Block-1000/#00-11 2.509Mi ± 0% 2.971Mi ± 0% +18.40% (p=0.002 n=6) DeduplicateFilter_Filter/Block-1000/#1-11 7.955Mi ± 0% 9.033Mi ± 0% +13.54% (p=0.002 n=6) DeduplicateFilter_Filter/Block-10000/#00-11 48.80Mi ± 0% 58.87Mi ± 0% +20.64% (p=0.002 n=6) DeduplicateFilter_Filter/Block-10000/#1-11 146.4Mi ± 0% 176.6Mi ± 0% +20.64% (p=0.002 n=6) geomean 1.606Mi 1.929Mi +20.13% │ before.txt │ after-with-bloom-filter-and-without-list.txt │ │ allocs/op │ allocs/op vs base │ DeduplicateFilter_Filter/Block-10/#00-11 63.00 ± 0% 64.00 ± 0% +1.59% (p=0.002 n=6) DeduplicateFilter_Filter/Block-10/#1-11 189.0 ± 0% 192.0 ± 0% +1.59% (p=0.002 n=6) DeduplicateFilter_Filter/Block-100/#00-11 522.0 ± 0% 523.0 ± 0% +0.19% (p=0.002 n=6) DeduplicateFilter_Filter/Block-100/#1-11 1.567k ± 0% 1.570k ± 0% +0.19% (p=0.002 n=6) DeduplicateFilter_Filter/Block-1000/#00-11 5.162k ± 0% 5.069k ± 0% -1.80% (p=0.002 n=6) DeduplicateFilter_Filter/Block-1000/#1-11 16.37k ± 0% 15.41k ± 0% -5.83% (p=0.002 n=6) DeduplicateFilter_Filter/Block-10000/#00-11 100.2k ± 0% 100.2k ± 0% +0.00% (p=0.002 n=6) DeduplicateFilter_Filter/Block-10000/#1-11 300.6k ± 0% 300.6k ± 0% +0.00% (p=0.002 n=6) geomean 3.542k 3.523k -0.53% ``` #### Which issue(s) this PR fixes or relates to N/A #### Checklist - [ ] Tests updated. - [ ] Documentation added. - [ ] `CHANGELOG.md` updated - the order of entries should be `[CHANGE]`, `[FEATURE]`, `[ENHANCEMENT]`, `[BUGFIX]`. If changelog entry is not needed, please add the `changelog-not-needed` label to the PR. - [ ] [`about-versioning.md`](https://github.com/grafana/mimir/blob/main/docs/sources/mimir/configure/about-versioning.md) updated with experimental features. --------- Signed-off-by: Marco Pracucci <[email protected]> Signed-off-by: Oleg Zaytsev <[email protected]> Co-authored-by: Oleg Zaytsev <[email protected]> Co-authored-by: Taylor C <[email protected]>
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 this PR does:
As first step after the forking, we should remove the deploy and deploy_website from CI. For now, CI should be just used to run tests.
Which issue(s) this PR fixes:
N/A
Checklist
CHANGELOG.mdupdated - the order of entries should be[CHANGE],[FEATURE],[ENHANCEMENT],[BUGFIX]