lib/promrelabel: use regexutil.PromRegex for regex matching in actions labeldrop,labelkeep,drop and keep

This makes possible optimizing additional cases inside regexutil.PromRegex

lib/promrelabel/config.go

@@ -8,6 +8,7 @@ import (
 
 	"github.com/VictoriaMetrics/VictoriaMetrics/lib/envtemplate"
 	"github.com/VictoriaMetrics/VictoriaMetrics/lib/fs"
+	"github.com/VictoriaMetrics/VictoriaMetrics/lib/logger"
 	"github.com/VictoriaMetrics/VictoriaMetrics/lib/regexutil"
 	"gopkg.in/yaml.v2"
 )
@@ -189,6 +190,13 @@ func ParseRelabelConfigs(rcs []RelabelConfig, relabelDebug bool) (*ParsedConfigs
 var (
 	defaultOriginalRegexForRelabelConfig = regexp.MustCompile(".*")
 	defaultRegexForRelabelConfig         = regexp.MustCompile("^(.*)$")
+	defaultPromRegex                     = func() *regexutil.PromRegex {
+		pr, err := regexutil.NewPromRegex(".*")
+		if err != nil {
+			panic(fmt.Errorf("BUG: unexpected error: %s", err))
+		}
+		return pr
+	}()
 )
 
 func parseRelabelConfig(rc *RelabelConfig) (*parsedRelabelConfig, error) {
@@ -202,9 +210,9 @@ func parseRelabelConfig(rc *RelabelConfig) (*parsedRelabelConfig, error) {
 		action = "replace"
 	}
 	targetLabel := rc.TargetLabel
-	regexCompiled := defaultRegexForRelabelConfig
+	regexAnchored := defaultRegexForRelabelConfig
 	regexOriginalCompiled := defaultOriginalRegexForRelabelConfig
-	var regexOrValues []string
+	promRegex := defaultPromRegex
 	if rc.Regex != nil && !isDefaultRegex(rc.Regex.S) {
 		regex := rc.Regex.S
 		regexOrig := regex
@@ -216,13 +224,16 @@ func parseRelabelConfig(rc *RelabelConfig) (*parsedRelabelConfig, error) {
 		if err != nil {
 			return nil, fmt.Errorf("cannot parse `regex` %q: %w", regex, err)
 		}
-		regexCompiled = re
+		regexAnchored = re
 		reOriginal, err := regexp.Compile(regexOrig)
 		if err != nil {
 			return nil, fmt.Errorf("cannot parse `regex` %q: %w", regexOrig, err)
 		}
 		regexOriginalCompiled = reOriginal
-		regexOrValues = regexutil.GetOrValues(regexOrig)
+		promRegex, err = regexutil.NewPromRegex(regexOrig)
+		if err != nil {
+			logger.Panicf("BUG: cannot parse already parsed regex %q: %s", regexOrig, err)
+		}
 	}
 	modulus := rc.Modulus
 	replacement := "$1"
@@ -335,20 +346,20 @@ func parseRelabelConfig(rc *RelabelConfig) (*parsedRelabelConfig, error) {
 		}
 	}
 	return &parsedRelabelConfig{
-		SourceLabels: sourceLabels,
+		SourceLabels:  sourceLabels,
-		Separator:    separator,
+		Separator:     separator,
-		TargetLabel:  targetLabel,
+		TargetLabel:   targetLabel,
-		Regex:        regexCompiled,
+		RegexAnchored: regexAnchored,
-		Modulus:      modulus,
+		Modulus:       modulus,
-		Replacement:  replacement,
+		Replacement:   replacement,
-		Action:       action,
+		Action:        action,
-		If:           rc.If,
+		If:            rc.If,
 
 		graphiteMatchTemplate: graphiteMatchTemplate,
 		graphiteLabelRules:    graphiteLabelRules,
 
+		regex:         promRegex,
 		regexOriginal: regexOriginalCompiled,
-		regexOrValues: regexOrValues,
 
 		hasCaptureGroupInTargetLabel:   strings.Contains(targetLabel, "$"),
 		hasCaptureGroupInReplacement:   strings.Contains(replacement, "$"),

lib/promrelabel/config_test.go

@@ -126,7 +126,7 @@ func TestParsedConfigsString(t *testing.T) {
 			TargetLabel:  "foo",
 			SourceLabels: []string{"aaa"},
 		},
-	}, "[SourceLabels=[aaa], Separator=;, TargetLabel=foo, Regex=^(.*)$, Modulus=0, Replacement=$1, Action=replace, If=, "+
+	}, "[SourceLabels=[aaa], Separator=;, TargetLabel=foo, Regex=.*, Modulus=0, Replacement=$1, Action=replace, If=, "+
 		"graphiteMatchTemplate=<nil>, graphiteLabelRules=[]], relabelDebug=false")
 	var ie IfExpression
 	if err := ie.Parse("{foo=~'bar'}"); err != nil {
@@ -141,7 +141,7 @@ func TestParsedConfigsString(t *testing.T) {
 			},
 			If: &ie,
 		},
-	}, "[SourceLabels=[], Separator=;, TargetLabel=, Regex=^(.*)$, Modulus=0, Replacement=$1, Action=graphite, If={foo=~'bar'}, "+
+	}, "[SourceLabels=[], Separator=;, TargetLabel=, Regex=.*, Modulus=0, Replacement=$1, Action=graphite, If={foo=~'bar'}, "+
 		"graphiteMatchTemplate=foo.*.bar, graphiteLabelRules=[replaceTemplate=$1-zz, targetLabel=job]], relabelDebug=false")
 	f([]RelabelConfig{
 		{
@@ -150,7 +150,7 @@ func TestParsedConfigsString(t *testing.T) {
 			TargetLabel:  "x",
 			If:           &ie,
 		},
-	}, "[SourceLabels=[foo bar], Separator=;, TargetLabel=x, Regex=^(.*)$, Modulus=0, Replacement=$1, Action=replace, If={foo=~'bar'}, "+
+	}, "[SourceLabels=[foo bar], Separator=;, TargetLabel=x, Regex=.*, Modulus=0, Replacement=$1, Action=replace, If={foo=~'bar'}, "+
 		"graphiteMatchTemplate=<nil>, graphiteLabelRules=[]], relabelDebug=false")
 }
 
@@ -174,13 +174,14 @@ func TestParseRelabelConfigsSuccess(t *testing.T) {
 	}, &ParsedConfigs{
 		prcs: []*parsedRelabelConfig{
 			{
-				SourceLabels: []string{"foo", "bar"},
+				SourceLabels:  []string{"foo", "bar"},
-				Separator:    ";",
+				Separator:     ";",
-				TargetLabel:  "xxx",
+				TargetLabel:   "xxx",
-				Regex:        defaultRegexForRelabelConfig,
+				RegexAnchored: defaultRegexForRelabelConfig,
-				Replacement:  "$1",
+				Replacement:   "$1",
-				Action:       "replace",
+				Action:        "replace",
 
+				regex:                        defaultPromRegex,
 				regexOriginal:                defaultOriginalRegexForRelabelConfig,
 				hasCaptureGroupInReplacement: true,
 			},

lib/promrelabel/relabel.go

@@ -9,6 +9,7 @@ import (
 	"github.com/VictoriaMetrics/VictoriaMetrics/lib/bytesutil"
 	"github.com/VictoriaMetrics/VictoriaMetrics/lib/logger"
 	"github.com/VictoriaMetrics/VictoriaMetrics/lib/prompbmarshal"
+	"github.com/VictoriaMetrics/VictoriaMetrics/lib/regexutil"
 	"github.com/cespare/xxhash/v2"
 )
 
@@ -16,20 +17,20 @@ import (
 //
 // See https://prometheus.io/docs/prometheus/latest/configuration/configuration/#relabel_config
 type parsedRelabelConfig struct {
-	SourceLabels []string
+	SourceLabels  []string
-	Separator    string
+	Separator     string
-	TargetLabel  string
+	TargetLabel   string
-	Regex        *regexp.Regexp
+	RegexAnchored *regexp.Regexp
-	Modulus      uint64
+	Modulus       uint64
-	Replacement  string
+	Replacement   string
-	Action       string
+	Action        string
-	If           *IfExpression
+	If            *IfExpression
 
 	graphiteMatchTemplate *graphiteMatchTemplate
 	graphiteLabelRules    []graphiteLabelRule
 
+	regex         *regexutil.PromRegex
 	regexOriginal *regexp.Regexp
-	regexOrValues []string
 
 	hasCaptureGroupInTargetLabel   bool
 	hasCaptureGroupInReplacement   bool
@@ -39,7 +40,8 @@ type parsedRelabelConfig struct {
 // String returns human-readable representation for prc.
 func (prc *parsedRelabelConfig) String() string {
 	return fmt.Sprintf("SourceLabels=%s, Separator=%s, TargetLabel=%s, Regex=%s, Modulus=%d, Replacement=%s, Action=%s, If=%s, graphiteMatchTemplate=%s, graphiteLabelRules=%s",
-		prc.SourceLabels, prc.Separator, prc.TargetLabel, prc.Regex, prc.Modulus, prc.Replacement, prc.Action, prc.If, prc.graphiteMatchTemplate, prc.graphiteLabelRules)
+		prc.SourceLabels, prc.Separator, prc.TargetLabel, prc.regexOriginal, prc.Modulus, prc.Replacement,
+		prc.Action, prc.If, prc.graphiteMatchTemplate, prc.graphiteLabelRules)
 }
 
 // Apply applies pcs to labels starting from the labelsOffset.
@@ -183,7 +185,7 @@ func (prc *parsedRelabelConfig) apply(labels []prompbmarshal.Label, labelsOffset
 			replacement = string(bb.B)
 		}
 		bb.B = concatLabelValues(bb.B[:0], src, prc.SourceLabels, prc.Separator)
-		if prc.Regex == defaultRegexForRelabelConfig && !prc.hasCaptureGroupInTargetLabel {
+		if prc.RegexAnchored == defaultRegexForRelabelConfig && !prc.hasCaptureGroupInTargetLabel {
 			if replacement == "$1" {
 				// Fast path for the rule that copies source label values to destination:
 				// - source_labels: [...]
@@ -201,7 +203,7 @@ func (prc *parsedRelabelConfig) apply(labels []prompbmarshal.Label, labelsOffset
 				return labels
 			}
 		}
-		match := prc.Regex.FindSubmatchIndex(bb.B)
+		match := prc.RegexAnchored.FindSubmatchIndex(bb.B)
 		if match == nil {
 			// Fast path - nothing to replace.
 			relabelBufPool.Put(bb)
@@ -253,7 +255,7 @@ func (prc *parsedRelabelConfig) apply(labels []prompbmarshal.Label, labelsOffset
 		return labels
 	case "keep":
 		// Keep the target if `source_labels` joined with `separator` match the `regex`.
-		if prc.Regex == defaultRegexForRelabelConfig {
+		if prc.RegexAnchored == defaultRegexForRelabelConfig {
 			// Fast path for the case with `if` and without explicitly set `regex`:
 			//
 			// - action: keep
@@ -263,7 +265,7 @@ func (prc *parsedRelabelConfig) apply(labels []prompbmarshal.Label, labelsOffset
 		}
 		bb := relabelBufPool.Get()
 		bb.B = concatLabelValues(bb.B[:0], src, prc.SourceLabels, prc.Separator)
-		keep := prc.matchString(bytesutil.ToUnsafeString(bb.B))
+		keep := prc.regex.MatchString(bytesutil.ToUnsafeString(bb.B))
 		relabelBufPool.Put(bb)
 		if !keep {
 			return labels[:labelsOffset]
@@ -271,7 +273,7 @@ func (prc *parsedRelabelConfig) apply(labels []prompbmarshal.Label, labelsOffset
 		return labels
 	case "drop":
 		// Drop the target if `source_labels` joined with `separator` don't match the `regex`.
-		if prc.Regex == defaultRegexForRelabelConfig {
+		if prc.RegexAnchored == defaultRegexForRelabelConfig {
 			// Fast path for the case with `if` and without explicitly set `regex`:
 			//
 			// - action: drop
@@ -281,7 +283,7 @@ func (prc *parsedRelabelConfig) apply(labels []prompbmarshal.Label, labelsOffset
 		}
 		bb := relabelBufPool.Get()
 		bb.B = concatLabelValues(bb.B[:0], src, prc.SourceLabels, prc.Separator)
-		drop := prc.matchString(bytesutil.ToUnsafeString(bb.B))
+		drop := prc.regex.MatchString(bytesutil.ToUnsafeString(bb.B))
 		relabelBufPool.Put(bb)
 		if drop {
 			return labels[:labelsOffset]
@@ -317,7 +319,7 @@ func (prc *parsedRelabelConfig) apply(labels []prompbmarshal.Label, labelsOffset
 		dst := labels[:labelsOffset]
 		for i := range src {
 			label := &src[i]
-			if !prc.matchString(label.Name) {
+			if !prc.regex.MatchString(label.Name) {
 				dst = append(dst, *label)
 			}
 		}
@@ -327,7 +329,7 @@ func (prc *parsedRelabelConfig) apply(labels []prompbmarshal.Label, labelsOffset
 		dst := labels[:labelsOffset]
 		for i := range src {
 			label := &src[i]
-			if prc.matchString(label.Name) {
+			if prc.regex.MatchString(label.Name) {
 				dst = append(dst, *label)
 			}
 		}
@@ -387,12 +389,12 @@ func (prc *parsedRelabelConfig) replaceFullString(s, replacement string, hasCapt
 		}
 	}
 	// Slow path - regexp processing
-	match := prc.Regex.FindStringSubmatchIndex(s)
+	match := prc.RegexAnchored.FindStringSubmatchIndex(s)
 	if match == nil {
 		return s, false
 	}
 	bb := relabelBufPool.Get()
-	bb.B = prc.Regex.ExpandString(bb.B[:0], replacement, s, match)
+	bb.B = prc.RegexAnchored.ExpandString(bb.B[:0], replacement, s, match)
 	result := string(bb.B)
 	relabelBufPool.Put(bb)
 	return result, true
@@ -413,39 +415,9 @@ func (prc *parsedRelabelConfig) replaceStringSubmatches(s, replacement string, h
 	return re.ReplaceAllString(s, replacement), true
 }
 
-func (prc *parsedRelabelConfig) matchString(s string) bool {
-	if len(prc.regexOrValues) > 0 {
-		for _, orValue := range prc.regexOrValues {
-			if s == orValue {
-				return true
-			}
-		}
-		return false
-	}
-	prefix, complete := prc.regexOriginal.LiteralPrefix()
-	if complete {
-		return prefix == s
-	}
-	if !strings.HasPrefix(s, prefix) {
-		return false
-	}
-	reStr := prc.regexOriginal.String()
-	if strings.HasPrefix(reStr, prefix) {
-		// Fast path for `foo.*` and `bar.+` regexps
-		reSuffix := reStr[len(prefix):]
-		switch reSuffix {
-		case ".+", "(.+)":
-			return len(s) > len(prefix)
-		case ".*", "(.*)":
-			return true
-		}
-	}
-	return prc.Regex.MatchString(s)
-}
-
 func (prc *parsedRelabelConfig) expandCaptureGroups(template, source string, match []int) string {
 	bb := relabelBufPool.Get()
-	bb.B = prc.Regex.ExpandString(bb.B[:0], template, source, match)
+	bb.B = prc.RegexAnchored.ExpandString(bb.B[:0], template, source, match)
 	s := string(bb.B)
 	relabelBufPool.Put(bb)
 	return s

lib/promrelabel/relabel_test.go

@@ -728,12 +728,12 @@ func TestFillLabelReferences(t *testing.T) {
 	f(`{{bar}}-aa{{__name__}}.{{bar}}{{non-existing-label}}`, `foo{bar="baz"}`, `baz-aafoo.baz`)
 }
 
-func TestRegexpMatchStringSuccess(t *testing.T) {
+func TestRegexMatchStringSuccess(t *testing.T) {
 	f := func(pattern, s string) {
 		t.Helper()
 		prc := newTestRegexRelabelConfig(pattern)
-		if !prc.matchString(s) {
+		if !prc.regex.MatchString(s) {
-			t.Fatalf("unexpected matchString(%q) result; got false; want true", s)
+			t.Fatalf("unexpected MatchString(%q) result; got false; want true", s)
 		}
 	}
 	f("", "")
@@ -753,8 +753,8 @@ func TestRegexpMatchStringFailure(t *testing.T) {
 	f := func(pattern, s string) {
 		t.Helper()
 		prc := newTestRegexRelabelConfig(pattern)
-		if prc.matchString(s) {
+		if prc.regex.MatchString(s) {
-			t.Fatalf("unexpected matchString(%q) result; got true; want false", s)
+			t.Fatalf("unexpected MatchString(%q) result; got true; want false", s)
 		}
 	}
 	f("", "foo")

lib/promrelabel/relabel_timing_test.go

@@ -16,7 +16,7 @@ func BenchmarkMatchRegexPrefixDotPlusMatchOptimized(b *testing.B) {
 	b.SetBytes(1)
 	b.RunParallel(func(pb *testing.PB) {
 		for pb.Next() {
-			if !prc.matchString(s) {
+			if !prc.regex.MatchString(s) {
 				panic(fmt.Errorf("unexpected string mismatch for pattern=%q, s=%q", pattern, s))
 			}
 		}
@@ -46,7 +46,7 @@ func BenchmarkMatchRegexPrefixDotPlusMismatchOptimized(b *testing.B) {
 	b.SetBytes(1)
 	b.RunParallel(func(pb *testing.PB) {
 		for pb.Next() {
-			if prc.matchString(s) {
+			if prc.regex.MatchString(s) {
 				panic(fmt.Errorf("unexpected string match for pattern=%q, s=%q", pattern, s))
 			}
 		}
@@ -76,7 +76,7 @@ func BenchmarkMatchRegexPrefixDotStarMatchOptimized(b *testing.B) {
 	b.SetBytes(1)
 	b.RunParallel(func(pb *testing.PB) {
 		for pb.Next() {
-			if !prc.matchString(s) {
+			if !prc.regex.MatchString(s) {
 				panic(fmt.Errorf("unexpected string mismatch for pattern=%q, s=%q", pattern, s))
 			}
 		}
@@ -106,7 +106,7 @@ func BenchmarkMatchRegexPrefixDotStarMismatchOptimized(b *testing.B) {
 	b.SetBytes(1)
 	b.RunParallel(func(pb *testing.PB) {
 		for pb.Next() {
-			if prc.matchString(s) {
+			if prc.regex.MatchString(s) {
 				panic(fmt.Errorf("unexpected string match for pattern=%q, s=%q", pattern, s))
 			}
 		}
@@ -136,7 +136,7 @@ func BenchmarkMatchRegexSingleValueMatchOptimized(b *testing.B) {
 	b.SetBytes(1)
 	b.RunParallel(func(pb *testing.PB) {
 		for pb.Next() {
-			if !prc.matchString(s) {
+			if !prc.regex.MatchString(s) {
 				panic(fmt.Errorf("unexpected string mismatch for pattern=%q, s=%q", pattern, s))
 			}
 		}
@@ -166,7 +166,7 @@ func BenchmarkMatchRegexSingleValueMismatchOptimized(b *testing.B) {
 	b.SetBytes(1)
 	b.RunParallel(func(pb *testing.PB) {
 		for pb.Next() {
-			if prc.matchString(s) {
+			if prc.regex.MatchString(s) {
 				panic(fmt.Errorf("unexpected string match for pattern=%q, s=%q", pattern, s))
 			}
 		}
@@ -196,7 +196,7 @@ func BenchmarkMatchRegexOrValuesMatchOptimized(b *testing.B) {
 	b.SetBytes(1)
 	b.RunParallel(func(pb *testing.PB) {
 		for pb.Next() {
-			if !prc.matchString(s) {
+			if !prc.regex.MatchString(s) {
 				panic(fmt.Errorf("unexpected string mismatch for pattern=%q, s=%q", pattern, s))
 			}
 		}
@@ -226,7 +226,7 @@ func BenchmarkMatchRegexOrValuesMismatchOptimized(b *testing.B) {
 	b.SetBytes(1)
 	b.RunParallel(func(pb *testing.PB) {
 		for pb.Next() {
-			if prc.matchString(s) {
+			if prc.regex.MatchString(s) {
 				panic(fmt.Errorf("unexpected string match for pattern=%q, s=%q", pattern, s))
 			}
 		}

lib/regexutil/promregex.go

@@ -16,7 +16,7 @@ import (
 type PromRegex struct {
 	// prefix contains literal prefix for regex.
 	// For example, prefix="foo" for regex="foo(a|b)"
-	prefix   string
+	prefix string
 
 	// Suffix contains regex suffix left after removing the prefix.
 	// For example, suffix="a|b" for regex="foo(a|b)"
@@ -51,12 +51,12 @@ func NewPromRegex(expr string) (*PromRegex, error) {
 	suffixExpr := "^(?:" + suffix + ")$"
 	reSuffix := regexp.MustCompile(suffixExpr)
 	pr := &PromRegex{
-		prefix:   prefix,
+		prefix:        prefix,
-		suffix: suffix,
+		suffix:        suffix,
 		substrDotStar: substrDotStar,
 		substrDotPlus: substrDotPlus,
-		orValues: orValues,
+		orValues:      orValues,
-		reSuffix:       reSuffix,
+		reSuffix:      reSuffix,
 	}
 	return pr, nil
 }
@@ -87,7 +87,7 @@ func (pr *PromRegex) MatchString(s string) bool {
 	if pr.substrDotPlus != "" {
 		// Fast path - pr contains ".+someText.+"
 		n := strings.Index(s, pr.substrDotPlus)
-		return n > 0 && n + len(pr.substrDotPlus) < len(s)
+		return n > 0 && n+len(pr.substrDotPlus) < len(s)
 	}
 	switch pr.suffix {
 	case ".*":
@@ -116,4 +116,3 @@ func getSubstringLiteral(expr, prefixSuffix string) string {
 	}
 	return prefix
 }
-

lib/regexutil/promregex_timing_test.go

@@ -19,6 +19,15 @@ func BenchmarkPromRegexMatchString(b *testing.B) {
 	b.Run("unpotimized-prefix-mismatch", func(b *testing.B) {
 		benchmarkPromRegexMatchString(b, "foo(bar.*|baz)", "zfoobarz", false)
 	})
+	b.Run("dot-star-match", func(b *testing.B) {
+		benchmarkPromRegexMatchString(b, ".*", "foo", true)
+	})
+	b.Run("dot-plus-match", func(b *testing.B) {
+		benchmarkPromRegexMatchString(b, ".+", "foo", true)
+	})
+	b.Run("dot-plus-mismatch", func(b *testing.B) {
+		benchmarkPromRegexMatchString(b, ".+", "", false)
+	})
 	b.Run("literal-match", func(b *testing.B) {
 		benchmarkPromRegexMatchString(b, "foo", "foo", true)
 	})