feat: 优化指标名校验性能 --story=119194100 (#486)

pkg/collector/VERSION

@@ -1 +1 @@
-0.65.x
+0.66.x

pkg/collector/internal/utils/normalize.go

@@ -17,3 +17,17 @@ import (
 func NormalizeName(s string) string {
 	return strings.Join(strings.FieldsFunc(s, func(r rune) bool { return !unicode.IsLetter(r) && !unicode.IsDigit(r) && r != '_' && r != ':' }), "_")
 }
+
+// IsNameNormalized 判断是否为合法指标名称
+// 此方式会比 regexp 方式要更快 参见 benchmark
+func IsNameNormalized(s string) bool {
+	if len(s) == 0 {
+		return false
+	}
+	for i, b := range s {
+		if !((b >= 'a' && b <= 'z') || (b >= 'A' && b <= 'Z') || b == '_' || (b >= '0' && b <= '9' && i > 0)) {
+			return false
+		}
+	}
+	return true
+}

pkg/collector/internal/utils/normalize_test.go

@@ -10,6 +10,7 @@
 package utils
 
 import (
+	"regexp"
 	"testing"
 
 	"github.com/stretchr/testify/assert"
@@ -56,3 +57,47 @@ func TestNormalize(t *testing.T) {
 		assert.Equal(t, c.Output, NormalizeName(c.Input))
 	}
 }
+
+func benchmarkIsNormalized(b *testing.B, f func(string) bool) {
+	type Case struct {
+		Input     string
+		Validated bool
+	}
+
+	cases := []Case{
+		{Input: "foo.bar", Validated: false},
+		{Input: "foo.bar.zzz", Validated: false},
+		{Input: "foo.bar..", Validated: false},
+		{Input: "TestApp_HelloGo_HelloGoObjAdapter_connectRate", Validated: true},
+		{Input: "TestApp_HelloGo_HelloGoObjAdapter.connectRate", Validated: false},
+		{Input: "TestApp.HelloGo.exception_single_log_more_than_3M", Validated: false},
+		{Input: "TestApp_HelloGo_exception_single_log_more_than_3M", Validated: true},
+		{Input: "TestApp.HelloGo.asyncqueue0", Validated: false},
+		{Input: "Exception-Log", Validated: false},
+		{Input: "┓(-´∀`-)┏", Validated: false},
+	}
+
+	b.ReportAllocs()
+	b.RunParallel(func(pb *testing.PB) {
+		for pb.Next() {
+			for _, c := range cases {
+				ok := f(c.Input)
+				if c.Validated != ok {
+					b.Errorf("input=(%v), want '%v' but go '%v'", c.Input, c.Validated, ok)
+				}
+			}
+		}
+	})
+}
+
+func BenchmarkIsNormalizedFast(b *testing.B) {
+	benchmarkIsNormalized(b, IsNameNormalized)
+}
+
+func BenchmarkIsNormalizedRegex(b *testing.B) {
+	namePattern := regexp.MustCompile("^[a-zA-Z_][a-zA-Z0-9_]*$")
+	f := func(s string) bool {
+		return namePattern.MatchString(s)
+	}
+	benchmarkIsNormalized(b, f)
+}

pkg/collector/processor/proxyvalidator/validator.go

@@ -10,13 +10,13 @@
 package proxyvalidator
 
 import (
-	"regexp"
 	"time"
 
 	"github.com/pkg/errors"
 	"github.com/spf13/cast"
 
 	"github.com/TencentBlueKing/bkmonitor-datalink/pkg/collector/define"
+	"github.com/TencentBlueKing/bkmonitor-datalink/pkg/collector/internal/utils"
 )
 
 const (
@@ -28,8 +28,6 @@ type Validator interface {
 	Validate(*define.ProxyData) error
 }
 
-var namePattern = regexp.MustCompile("^[a-zA-Z_][a-zA-Z0-9_]*$")
-
 type noneValidator struct{}
 
 func (noneValidator) Validate(*define.ProxyData) error {
@@ -39,10 +37,9 @@ func (noneValidator) Validate(*define.ProxyData) error {
 type nameValidator struct{}
 
 func (nv *nameValidator) Validate(s string) error {
-	if !namePattern.MatchString(s) {
+	if !utils.IsNameNormalized(s) {
 		return errors.Errorf("name '%s' required match regex [^[a-zA-Z_][a-zA-Z0-9_]*$]", s)
 	}
-
 	return nil
 }