switch to prometheus name escaping function

open-telemetry · Sep 9, 2024 · ff43372 · ff43372
1 parent ff554f3
commit ff43372
Show file tree

Hide file tree

Showing 4 changed files with 10 additions and 100 deletions.
diff --git a/exporters/prometheus/config.go b/exporters/prometheus/config.go
@@ -7,6 +7,7 @@ import (
 	"strings"
 
 	"github.com/prometheus/client_golang/prometheus"
+	"github.com/prometheus/common/model"
 
 	"go.opentelemetry.io/otel/attribute"
 	"go.opentelemetry.io/otel/sdk/metric"
@@ -131,7 +132,10 @@ func WithoutScopeInfo() Option {
 // have special behavior based on their name.
 func WithNamespace(ns string) Option {
 	return optionFunc(func(cfg config) config {
-		ns = sanitizeName(ns)
+		if model.NameValidationScheme != model.UTF8Validation {
+			// Only sanitize if prometheus does not support UTF-8.
+			ns = model.EscapeName(ns, model.NameEscapingScheme)
+		}
 		if !strings.HasSuffix(ns, "_") {
 			// namespace and metric names should be separated with an underscore,
 			// adds a trailing underscore if there is not one already.

diff --git a/exporters/prometheus/exporter.go b/exporters/prometheus/exporter.go
@@ -11,8 +11,6 @@ import (
 	"slices"
 	"strings"
 	"sync"
-	"unicode"
-	"unicode/utf8"
 
 	"github.com/prometheus/client_golang/prometheus"
 	dto "github.com/prometheus/client_model/go"
@@ -318,7 +316,7 @@ func getAttrs(attrs attribute.Set, ks, vs [2]string, resourceKV keyVals) ([]stri
 		keysMap := make(map[string][]string)
 		for itr.Next() {
 			kv := itr.Attribute()
-			key := strings.Map(sanitizeRune, string(kv.Key))
+			key := model.EscapeName(string(kv.Key), model.NameEscapingScheme)
 			if _, ok := keysMap[key]; !ok {
 				keysMap[key] = []string{kv.Value.Emit()}
 			} else {
@@ -358,13 +356,6 @@ func createScopeInfoMetric(scope instrumentation.Scope) (prometheus.Metric, erro
 	return prometheus.NewConstMetric(desc, prometheus.GaugeValue, float64(1), scope.Name, scope.Version)
 }
 
-func sanitizeRune(r rune) rune {
-	if unicode.IsLetter(r) || unicode.IsDigit(r) || r == ':' || r == '_' {
-		return r
-	}
-	return '_'
-}
-
 var unitSuffixes = map[string]string{
 	// Time
 	"d":   "_days",
@@ -406,7 +397,7 @@ func (c *collector) getName(m metricdata.Metrics, typ *dto.MetricType) string {
 	name := m.Name
 	if model.NameValidationScheme != model.UTF8Validation {
 		// Only sanitize if prometheus does not support UTF-8.
-		name = sanitizeName(m.Name)
+		name = model.EscapeName(name, model.NameEscapingScheme)
 	}
 	addCounterSuffix := !c.withoutCounterSuffixes && *typ == dto.MetricType_COUNTER
 	if addCounterSuffix {
@@ -426,59 +417,6 @@ func (c *collector) getName(m metricdata.Metrics, typ *dto.MetricType) string {
 	return name
 }
 
-func sanitizeName(n string) string {
-	// This algorithm is based on strings.Map from Go 1.19.
-	const replacement = '_'
-
-	valid := func(i int, r rune) bool {
-		// Taken from
-		// https://github.com/prometheus/common/blob/dfbc25bd00225c70aca0d94c3c4bb7744f28ace0/model/metric.go#L92-L102
-		if (r >= 'a' && r <= 'z') || (r >= 'A' && r <= 'Z') || r == '_' || r == ':' || (r >= '0' && r <= '9' && i > 0) {
-			return true
-		}
-		return false
-	}
-
-	// This output buffer b is initialized on demand, the first time a
-	// character needs to be replaced.
-	var b strings.Builder
-	for i, c := range n {
-		if valid(i, c) {
-			continue
-		}
-
-		if i == 0 && c >= '0' && c <= '9' {
-			// Prefix leading number with replacement character.
-			b.Grow(len(n) + 1)
-			_ = b.WriteByte(byte(replacement))
-			break
-		}
-		b.Grow(len(n))
-		_, _ = b.WriteString(n[:i])
-		_ = b.WriteByte(byte(replacement))
-		width := utf8.RuneLen(c)
-		n = n[i+width:]
-		break
-	}
-
-	// Fast path for unchanged input.
-	if b.Cap() == 0 { // b.Grow was not called above.
-		return n
-	}
-
-	for _, c := range n {
-		// Due to inlining, it is more performant to invoke WriteByte rather then
-		// WriteRune.
-		if valid(1, c) { // We are guaranteed to not be at the start.
-			_ = b.WriteByte(byte(c))
-		} else {
-			_ = b.WriteByte(byte(replacement))
-		}
-	}
-
-	return b.String()
-}
-
 func (c *collector) metricType(m metricdata.Metrics) *dto.MetricType {
 	switch v := m.Data.(type) {
 	case metricdata.Histogram[int64], metricdata.Histogram[float64]:

diff --git a/exporters/prometheus/exporter_test.go b/exporters/prometheus/exporter_test.go
@@ -437,11 +437,9 @@ func TestPrometheusExporter(t *testing.T) {
 		t.Run(tc.name, func(t *testing.T) {
 			if tc.enableUTF8 {
 				model.NameValidationScheme = model.UTF8Validation
-				model.NameEscapingScheme = model.NoEscaping
 				defer func() {
 					// Reset to defaults
 					model.NameValidationScheme = model.LegacyValidation
-					model.NameEscapingScheme = model.ValueEncodingEscaping
 				}()
 			}
 			ctx := context.Background()
@@ -493,36 +491,6 @@ func TestPrometheusExporter(t *testing.T) {
 	}
 }
 
-func TestSantitizeName(t *testing.T) {
-	tests := []struct {
-		input string
-		want  string
-	}{
-		{"name€_with_4_width_rune", "name__with_4_width_rune"},
-		{"`", "_"},
-		{
-			`! "#$%&'()*+,-./0123456789:;<=>?@ABCDEFGHIJKLMNOPQRSTUVWKYZ[]\^_abcdefghijklmnopqrstuvwkyz{|}~`,
-			`________________0123456789:______ABCDEFGHIJKLMNOPQRSTUVWKYZ_____abcdefghijklmnopqrstuvwkyz____`,
-		},
-
-		// Test cases taken from
-		// https://github.com/prometheus/common/blob/dfbc25bd00225c70aca0d94c3c4bb7744f28ace0/model/metric_test.go#L85-L136
-		{"Avalid_23name", "Avalid_23name"},
-		{"_Avalid_23name", "_Avalid_23name"},
-		{"1valid_23name", "_1valid_23name"},
-		{"avalid_23name", "avalid_23name"},
-		{"Ava:lid_23name", "Ava:lid_23name"},
-		{"a lid_23name", "a_lid_23name"},
-		{":leading_colon", ":leading_colon"},
-		{"colon:in:the:middle", "colon:in:the:middle"},
-		{"", ""},
-	}
-
-	for _, test := range tests {
-		require.Equalf(t, test.want, sanitizeName(test.input), "input: %q", test.input)
-	}
-}
-
 func TestMultiScopes(t *testing.T) {
 	ctx := context.Background()
 	registry := prometheus.NewRegistry()

diff --git a/exporters/prometheus/testdata/sanitized_names.txt b/exporters/prometheus/testdata/sanitized_names.txt
@@ -1,9 +1,9 @@
 # HELP bar a fun little gauge
 # TYPE bar gauge
 bar{A="B",C="D",otel_scope_name="testmeter",otel_scope_version="v0.1.0"} 75
-# HELP _0invalid_counter_name_total a counter with an invalid name
-# TYPE _0invalid_counter_name_total counter
-_0invalid_counter_name_total{A="B",C="D",otel_scope_name="testmeter",otel_scope_version="v0.1.0"} 100
+# HELP _invalid_counter_name_total a counter with an invalid name
+# TYPE _invalid_counter_name_total counter
+_invalid_counter_name_total{A="B",C="D",otel_scope_name="testmeter",otel_scope_version="v0.1.0"} 100
 # HELP invalid_gauge_name a gauge with an invalid name
 # TYPE invalid_gauge_name gauge
 invalid_gauge_name{A="B",C="D",otel_scope_name="testmeter",otel_scope_version="v0.1.0"} 100