Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

Optimize l=~".+" matcher #15474

Merged
merged 1 commit into from
Nov 27, 2024
Merged

Optimize l=~".+" matcher #15474

merged 1 commit into from
Nov 27, 2024

Conversation

colega
Copy link
Contributor

@colega colega commented Nov 27, 2024

Since dot is matching newline now, l=~".+" is "any non empty label value", and #14144 added a specific method in the index for that so we don't need to run the matcher on each one of the label values.

I only ran the relevant benchmarks:

goos: darwin
goarch: arm64
pkg: github.com/prometheus/prometheus/tsdb
cpu: Apple M3 Pro
                                                                     │   dp_main    │             dp_shortcut             │
                                                                     │    sec/op    │    sec/op     vs base               │
Querier/Head/PostingsForMatchers/i=~".+"-12                            9.203m ± 15%   4.989m ± 16%  -45.78% (p=0.002 n=6)
Querier/Head/PostingsForMatchers/i=~".+",j=~"X.+"-12                   9.130m ± 69%   4.813m ±  7%  -47.29% (p=0.002 n=6)
Querier/Head/PostingsForMatchers/n="1",i=~".+",j="foo"-12              8.374m ±  9%   4.925m ±  8%  -41.18% (p=0.002 n=6)
Querier/Head/PostingsForMatchers/n="1",i=~".+",i!="2",j="foo"-12       8.552m ± 21%   5.141m ± 10%  -39.89% (p=0.002 n=6)
Querier/Head/PostingsForMatchers/n="1",i=~".+",i!~"2.*",j="foo"-12     13.31m ± 15%   10.04m ±  9%  -24.56% (p=0.002 n=6)
Querier/Head/PostingsForMatchers/n="1",i=~".+",i!~".*2.*",j="foo"-12   18.65m ±  3%   15.62m ±  4%  -16.26% (p=0.002 n=6)
Querier/Head/PostingsForMatchers/n="X",i=~".+",i!~".*2.*",j="foo"-12   298.9n ±  3%   303.1n ± 11%        ~ (p=0.240 n=6)
geomean                                                                2.391m         1.617m        -32.37%

                                                                     │   dp_main    │              dp_shortcut              │
                                                                     │     B/op     │     B/op      vs base                 │
Querier/Head/PostingsForMatchers/i=~".+"-12                            12.23Mi ± 0%   10.70Mi ± 0%  -12.52% (p=0.002 n=6)
Querier/Head/PostingsForMatchers/i=~".+",j=~"X.+"-12                   12.23Mi ± 0%   10.70Mi ± 0%  -12.52% (p=0.002 n=6)
Querier/Head/PostingsForMatchers/n="1",i=~".+",j="foo"-12              12.23Mi ± 0%   10.70Mi ± 0%  -12.52% (p=0.002 n=6)
Querier/Head/PostingsForMatchers/n="1",i=~".+",i!="2",j="foo"-12       12.23Mi ± 0%   10.70Mi ± 0%  -12.52% (p=0.002 n=6)
Querier/Head/PostingsForMatchers/n="1",i=~".+",i!~"2.*",j="foo"-12     14.96Mi ± 0%   13.43Mi ± 0%  -10.24% (p=0.002 n=6)
Querier/Head/PostingsForMatchers/n="1",i=~".+",i!~".*2.*",j="foo"-12   18.14Mi ± 0%   16.61Mi ± 0%   -8.44% (p=0.002 n=6)
Querier/Head/PostingsForMatchers/n="X",i=~".+",i!~".*2.*",j="foo"-12     48.00 ± 0%     48.00 ± 0%        ~ (p=1.000 n=6) ¹
geomean                                                                2.234Mi        2.012Mi        -9.92%
¹ all samples are equal

                                                                     │  dp_main   │             dp_shortcut             │
                                                                     │ allocs/op  │ allocs/op   vs base                 │
Querier/Head/PostingsForMatchers/i=~".+"-12                            8.000 ± 0%   6.000 ± 0%  -25.00% (p=0.002 n=6)
Querier/Head/PostingsForMatchers/i=~".+",j=~"X.+"-12                   12.00 ± 0%   10.00 ± 0%  -16.67% (p=0.002 n=6)
Querier/Head/PostingsForMatchers/n="1",i=~".+",j="foo"-12              15.00 ± 0%   13.00 ± 0%  -13.33% (p=0.002 n=6)
Querier/Head/PostingsForMatchers/n="1",i=~".+",i!="2",j="foo"-12       20.00 ± 0%   18.00 ± 0%  -10.00% (p=0.002 n=6)
Querier/Head/PostingsForMatchers/n="1",i=~".+",i!~"2.*",j="foo"-12     92.00 ± 0%   90.00 ± 0%   -2.17% (p=0.002 n=6)
Querier/Head/PostingsForMatchers/n="1",i=~".+",i!~".*2.*",j="foo"-12   95.00 ± 0%   93.00 ± 0%   -2.11% (p=0.002 n=6)
Querier/Head/PostingsForMatchers/n="X",i=~".+",i!~".*2.*",j="foo"-12   3.000 ± 0%   3.000 ± 0%        ~ (p=1.000 n=6) ¹
geomean                                                                18.55        16.64       -10.31%
¹ all samples are equal

@colega
Optimize l=~".+" matcher
14e937a 
Since dot is matching newline now, `l=~".+"` is "any non empty label
value", and prometheus#14144 added a specific method in the index for that so we
don't need to run the matcher on each one of the label values.

Signed-off-by: Oleg Zaytsev <[email protected]>
@colega colega requested a review from jesusvazquez as a code owner November 27, 2024 11:08
aknuds1
aknuds1 approved these changes Nov 27, 2024
View reviewed changes
Copy link
Contributor

@aknuds1 aknuds1 left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

LGTM!

tsdb/querier.go
@@ -256,10 +256,23 @@ func PostingsForMatchers(ctx context.Context, ix IndexReader, ms ...*labels.Matc
// We already handled the case at the top of the function,
// and it is unexpected to get all postings again here.
return nil, errors.New("unexpected all postings")

Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Just a nit, but I'd prefer staying consistent and not adding blank lines between the switch branches.

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I added the lines to visually group the different cases instead of having a continuous chunk of code: the previous two cases handle the ~".*" matchers, and these two handle ~".+"

Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Yeah I realized. I still prefer consistency :) Makes for easy maintenance.

@aknuds1 aknuds1 merged commit 9ad93ba into prometheus:main Nov 27, 2024
26 checks passed
@charleskorn charleskorn mentioned this pull request Dec 1, 2024
Closed
@krajorama krajorama mentioned this pull request Dec 9, 2024
Merged
@yeya24 yeya24 mentioned this pull request Mar 1, 2025
Merged
2 tasks
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Reviewers

@aknuds1 aknuds1 aknuds1 approved these changes

@jesusvazquez jesusvazquez Awaiting requested review from jesusvazquez jesusvazquez is a code owner

Assignees
No one assigned
Labels
None yet
Projects
None yet
Milestone
No milestone
Development

Successfully merging this pull request may close these issues.
2 participants
@colega @aknuds1