Optimize wildcard matchers for .* and .+ (thanos-io#8131)

* optimize wildcard matchers for .* and .+ Signed-off-by: yeya24 <benye@amazon.com> * add changelog Signed-off-by: yeya24 <benye@amazon.com> --------- Signed-off-by: yeya24 <benye@amazon.com>
yeya24 · Mar 3, 2025 · c69f112 · c69f112
1 parent 4ba7d59
commit c69f112
Show file tree

Hide file tree

Showing 3 changed files with 78 additions and 0 deletions.
diff --git a/CHANGELOG.md b/CHANGELOG.md
@@ -23,6 +23,7 @@ We use *breaking :warning:* to mark changes that are not backward compatible (re
 - [#8000](https://github.com/thanos-io/thanos/pull/8000) Query: Bump promql-engine, pass partial response through options
 - [#7353](https://github.com/thanos-io/thanos/pull/7353) [#8045](https://github.com/thanos-io/thanos/pull/8045) Receiver/StoreGateway: Add `--matcher-cache-size` option to enable caching for regex matchers in series calls.
 - [#8017](https://github.com/thanos-io/thanos/pull/8017) Store Gateway: Use native histogram for binary reader load and download duration and fixed download duration metric. #8017
+- [#8131](https://github.com/thanos-io/thanos/pull/8131) Store Gateway: Optimize regex matchers for .* and .+. #8131
 
 ### Changed
 

diff --git a/pkg/store/bucket.go b/pkg/store/bucket.go
@@ -2958,6 +2958,16 @@ func matchersToPostingGroups(ctx context.Context, lvalsFn func(name string) ([]s
 
 // NOTE: Derived from tsdb.postingsForMatcher. index.Merge is equivalent to map duplication.
 func toPostingGroup(ctx context.Context, lvalsFn func(name string) ([]string, error), m *labels.Matcher) (*postingGroup, []string, error) {
+	// .* regexp matches any string.
+	if m.Type == labels.MatchRegexp && m.Value == ".*" {
+		return newPostingGroup(true, m.Name, nil, nil), nil, nil
+	}
+
+	// .* not regexp doesn't match any string.
+	if m.Type == labels.MatchNotRegexp && m.Value == ".*" {
+		return newPostingGroup(false, m.Name, nil, nil), nil, nil
+	}
+
 	// If the matcher selects an empty value, it selects all the series which don't
 	// have the label name set too. See: https://github.com/prometheus/prometheus/issues/3575
 	// and https://github.com/prometheus/prometheus/pull/3578#issuecomment-351653555.
@@ -2991,6 +3001,11 @@ func toPostingGroup(ctx context.Context, lvalsFn func(name string) ([]string, er
 			return newPostingGroup(true, m.Name, nil, vals), vals, nil
 		}
 
+		// .+ regexp matches any non-empty string: get postings for all label values and remove them.
+		if m.Type == labels.MatchNotRegexp && m.Value == ".+" {
+			return newPostingGroup(true, m.Name, nil, vals), vals, nil
+		}
+
 		for i, val := range vals {
 			if (i+1)%checkContextEveryNIterations == 0 && ctx.Err() != nil {
 				return nil, nil, ctx.Err()
@@ -3002,6 +3017,7 @@ func toPostingGroup(ctx context.Context, lvalsFn func(name string) ([]string, er
 
 		return newPostingGroup(true, m.Name, nil, toRemove), vals, nil
 	}
+
 	if m.Type == labels.MatchRegexp {
 		if vals := m.SetMatches(); len(vals) > 0 {
 			sort.Strings(vals)
@@ -3025,6 +3041,11 @@ func toPostingGroup(ctx context.Context, lvalsFn func(name string) ([]string, er
 		return newPostingGroup(false, m.Name, vals, nil), vals, nil
 	}
 
+	// .+ regexp matches any non-empty string: get postings for all label values.
+	if m.Type == labels.MatchRegexp && m.Value == ".+" {
+		return newPostingGroup(false, m.Name, vals, nil), vals, nil
+	}
+
 	var toAdd []string
 	for i, val := range vals {
 		if (i+1)%checkContextEveryNIterations == 0 && ctx.Err() != nil {

diff --git a/pkg/store/bucket_test.go b/pkg/store/bucket_test.go
@@ -3492,6 +3492,62 @@ func TestPostingGroupMerge(t *testing.T) {
 	}
 }
 
+func TestToPostingGroup(t *testing.T) {
+	ctx := context.Background()
+	for _, tc := range []struct {
+		name                  string
+		vals                  []string
+		matcher               *labels.Matcher
+		addAll                bool
+		addKeys               []string
+		removeKeys            []string
+		numberLabelValueCalls int
+	}{
+		{
+			name:                  "regexp .*",
+			matcher:               labels.MustNewMatcher(labels.MatchRegexp, labels.MetricName, ".*"),
+			addAll:                true,
+			numberLabelValueCalls: 0,
+		},
+		{
+			name:                  "not regexp .*",
+			matcher:               labels.MustNewMatcher(labels.MatchNotRegexp, labels.MetricName, ".*"),
+			addAll:                false,
+			numberLabelValueCalls: 0,
+		},
+		{
+			name:                  "regexp .+",
+			matcher:               labels.MustNewMatcher(labels.MatchRegexp, labels.MetricName, ".+"),
+			addAll:                false,
+			addKeys:               []string{"foo"},
+			vals:                  []string{"foo"},
+			numberLabelValueCalls: 1,
+		},
+		{
+			name:                  "not regexp .+",
+			matcher:               labels.MustNewMatcher(labels.MatchNotRegexp, labels.MetricName, ".+"),
+			addAll:                true,
+			removeKeys:            []string{"foo"},
+			vals:                  []string{"foo"},
+			numberLabelValueCalls: 1,
+		},
+	} {
+		t.Run(tc.name, func(t *testing.T) {
+			calls := 0
+			lvalsFn := func(name string) ([]string, error) {
+				calls++
+				return tc.vals, nil
+			}
+			pg, _, err := toPostingGroup(ctx, lvalsFn, tc.matcher)
+			testutil.Ok(t, err)
+			testutil.Equals(t, tc.addAll, pg.addAll)
+			testutil.Equals(t, tc.addKeys, pg.addKeys)
+			testutil.Equals(t, tc.removeKeys, pg.removeKeys)
+			testutil.Equals(t, tc.numberLabelValueCalls, calls)
+		})
+	}
+}
+
 // TestExpandedPostings is a test whether there is a race between multiple ExpandPostings() calls.
 func TestExpandedPostingsRace(t *testing.T) {
 	t.Parallel()