Store: big values set for bucket index benchmark

Signed-off-by: Xiaochao Dong (@damnever) <the.xcdong@gmail.com>
thanos-io · Aug 16, 2022 · ceee652 · ceee652
1 parent b0b87fa
commit ceee652
Showing 1 changed file with 22 additions and 5 deletions.
diff --git a/pkg/store/bucket_test.go b/pkg/store/bucket_test.go
@@ -17,6 +17,7 @@ import (
 	"regexp"
 	"sort"
 	"strconv"
+	"strings"
 	"sync"
 	"testing"
 	"time"
@@ -1067,14 +1068,16 @@ func appendTestData(t testing.TB, app storage.Appender, series int) {
 	}
 
 	series = series / 5
+	uniq := 0
 	for n := 0; n < 10; n++ {
 		for i := 0; i < series/10; i++ {
-			addSeries(labels.FromStrings("i", strconv.Itoa(i)+storetestutil.LabelLongSuffix, "n", strconv.Itoa(n)+storetestutil.LabelLongSuffix, "j", "foo"))
+			addSeries(labels.FromStrings("i", strconv.Itoa(i)+storetestutil.LabelLongSuffix, "n", strconv.Itoa(n)+storetestutil.LabelLongSuffix, "j", "foo", "uniq", fmt.Sprintf("%08d", uniq)))
 			// Have some series that won't be matched, to properly test inverted matches.
-			addSeries(labels.FromStrings("i", strconv.Itoa(i)+storetestutil.LabelLongSuffix, "n", strconv.Itoa(n)+storetestutil.LabelLongSuffix, "j", "bar"))
-			addSeries(labels.FromStrings("i", strconv.Itoa(i)+storetestutil.LabelLongSuffix, "n", "0_"+strconv.Itoa(n)+storetestutil.LabelLongSuffix, "j", "bar"))
-			addSeries(labels.FromStrings("i", strconv.Itoa(i)+storetestutil.LabelLongSuffix, "n", "1_"+strconv.Itoa(n)+storetestutil.LabelLongSuffix, "j", "bar"))
-			addSeries(labels.FromStrings("i", strconv.Itoa(i)+storetestutil.LabelLongSuffix, "n", "2_"+strconv.Itoa(n)+storetestutil.LabelLongSuffix, "j", "foo"))
+			addSeries(labels.FromStrings("i", strconv.Itoa(i)+storetestutil.LabelLongSuffix, "n", strconv.Itoa(n)+storetestutil.LabelLongSuffix, "j", "bar", "uniq", fmt.Sprintf("%08d", uniq+1)))
+			addSeries(labels.FromStrings("i", strconv.Itoa(i)+storetestutil.LabelLongSuffix, "n", "0_"+strconv.Itoa(n)+storetestutil.LabelLongSuffix, "j", "bar", "uniq", fmt.Sprintf("%08d", uniq+2)))
+			addSeries(labels.FromStrings("i", strconv.Itoa(i)+storetestutil.LabelLongSuffix, "n", "1_"+strconv.Itoa(n)+storetestutil.LabelLongSuffix, "j", "bar", "uniq", fmt.Sprintf("%08d", uniq+3)))
+			addSeries(labels.FromStrings("i", strconv.Itoa(i)+storetestutil.LabelLongSuffix, "n", "2_"+strconv.Itoa(n)+storetestutil.LabelLongSuffix, "j", "foo", "uniq", fmt.Sprintf("%08d", uniq+4)))
+			uniq += 5
 		}
 	}
 	testutil.Ok(t, app.Commit())
@@ -1115,6 +1118,19 @@ func benchmarkExpandedPostings(
 	iNot2 := labels.MustNewMatcher(labels.MatchNotEqual, "n", "2"+storetestutil.LabelLongSuffix)
 	iNot2Star := labels.MustNewMatcher(labels.MatchNotRegexp, "i", "^2.*$")
 	iRegexSet := labels.MustNewMatcher(labels.MatchRegexp, "i", "0"+storetestutil.LabelLongSuffix+"|1"+storetestutil.LabelLongSuffix+"|2"+storetestutil.LabelLongSuffix)
+	bigValueSetSize := series / 10
+	if bigValueSetSize > 50000 {
+		bigValueSetSize = 50000
+	}
+	bigValueSet := make([]string, 0, bigValueSetSize)
+	for i := 0; i < series; i += series / bigValueSetSize {
+		bigValueSet = append(bigValueSet, fmt.Sprintf("%08d", i))
+	}
+	bigValueSetSize = len(bigValueSet)
+	rand.New(rand.NewSource(time.Now().UnixNano())).Shuffle(len(bigValueSet), func(i, j int) {
+		bigValueSet[i], bigValueSet[j] = bigValueSet[j], bigValueSet[i]
+	})
+	iRegexBigValueSet := labels.MustNewMatcher(labels.MatchRegexp, "uniq", strings.Join(bigValueSet, "|"))
 
 	series = series / 5
 	cases := []struct {
@@ -1140,6 +1156,7 @@ func benchmarkExpandedPostings(
 		{`n="1",i=~".+",i!="2",j="foo"`, []*labels.Matcher{n1, iPlus, iNot2, jFoo}, int(float64(series) * 0.1)},
 		{`n="1",i=~".+",i!~"2.*",j="foo"`, []*labels.Matcher{n1, iPlus, iNot2Star, jFoo}, int(1 + float64(series)*0.088888)},
 		{`i=~"0|1|2"`, []*labels.Matcher{iRegexSet}, 150}, // 50 series for "1", 50 for "2" and 50 for "3".
+		{`uniq=~"9|random-shuffled-values|1"`, []*labels.Matcher{iRegexBigValueSet}, bigValueSetSize},
 	}
 
 	for _, c := range cases {