otlptranslator: Harmonize non-UTF8 sanitization w/ naming rules. (#15314)

* otlptranslator: Harmonize non-UTF8 sanitization w/ naming rules.

Harmonize non-UTF8 sanitization w/ Prometheus naming rules.

---------

Signed-off-by: Arve Knudsen <arve.knudsen@gmail.com>

CHANGELOG.md

@@ -13,6 +13,7 @@
 * [ENHANCEMENT] Scraping: support Created-Timestamp feature on native histograms. #14694
 * [BUGFIX] PromQL: Fix stddev+stdvar aggregations to always ignore native histograms. #14941
 * [BUGFIX] PromQL: Fix stddev+stdvar aggregations to treat Infinity consistently. #14941
+* [BUGFIX] OTLP receiver: Preserve colons when generating metric names in suffix adding mode (this mode is always enabled, unless one uses Prometheus as a library). #15251
 
 ## 3.0.0-beta.1 / 2024-10-09
 

storage/remote/otlptranslator/prometheus/normalize_label.go

@@ -19,6 +19,8 @@ package prometheus
 import (
 	"strings"
 	"unicode"
+
+	"github.com/prometheus/prometheus/util/strutil"
 )
 
 // Normalizes the specified label to follow Prometheus label names standard.
@@ -26,7 +28,6 @@ import (
 // See rules at https://prometheus.io/docs/concepts/data_model/#metric-names-and-labels.
 //
 // Labels that start with non-letter rune will be prefixed with "key_".
-//
 // An exception is made for double-underscores which are allowed.
 func NormalizeLabel(label string) string {
 	// Trivial case
@@ -34,8 +35,7 @@ func NormalizeLabel(label string) string {
 		return label
 	}
 
-	// Replace all non-alphanumeric runes with underscores
-	label = strings.Map(sanitizeRune, label)
+	label = strutil.SanitizeLabelName(label)
 
 	// If label starts with a number, prepend with "key_"
 	if unicode.IsDigit(rune(label[0])) {
@@ -46,11 +46,3 @@ func NormalizeLabel(label string) string {
 
 	return label
 }
-
-// Return '_' for anything non-alphanumeric.
-func sanitizeRune(r rune) rune {
-	if unicode.IsLower(r) || unicode.IsUpper(r) || unicode.IsDigit(r) {
-		return r
-	}
-	return '_'
-}

storage/remote/otlptranslator/prometheus/normalize_name.go

@@ -17,9 +17,12 @@
 package prometheus
 
 import (
+	"regexp"
+	"slices"
 	"strings"
 	"unicode"
 
+	"github.com/prometheus/prometheus/util/strutil"
 	"go.opentelemetry.io/collector/pdata/pmetric"
 )
 
@@ -84,24 +87,27 @@ var perUnitMap = map[string]string{
 //
 // See rules at https://prometheus.io/docs/concepts/data_model/#metric-names-and-labels,
 // https://prometheus.io/docs/practices/naming/#metric-and-label-naming
-// and https://github.com/open-telemetry/opentelemetry-specification/blob/v1.33.0/specification/compatibility/prometheus_and_openmetrics.md#otlp-metric-points-to-prometheus.
+// and https://github.com/open-telemetry/opentelemetry-specification/blob/v1.38.0/specification/compatibility/prometheus_and_openmetrics.md#otlp-metric-points-to-prometheus.
 func BuildCompliantName(metric pmetric.Metric, namespace string, addMetricSuffixes bool) string {
-	var metricName string
-
 	// Full normalization following standard Prometheus naming conventions
 	if addMetricSuffixes {
 		return normalizeName(metric, namespace)
 	}
 
-	// Simple case (no full normalization, no units, etc.), we simply trim out forbidden chars
-	metricName = RemovePromForbiddenRunes(metric.Name())
+	// Regexp for metric name characters that should be replaced with _.
+	invalidMetricCharRE := regexp.MustCompile(`[^a-zA-Z0-9:_]`)
+
+	// Simple case (no full normalization, no units, etc.).
+	metricName := strings.Join(strings.FieldsFunc(metric.Name(), func(r rune) bool {
+		return invalidMetricCharRE.MatchString(string(r))
+	}), "_")
 
 	// Namespace?
 	if namespace != "" {
 		return namespace + "_" + metricName
 	}
 
-	// Metric name starts with a digit? Prefix it with an underscore
+	// Metric name starts with a digit? Prefix it with an underscore.
 	if metricName != "" && unicode.IsDigit(rune(metricName[0])) {
 		metricName = "_" + metricName
 	}
@@ -109,12 +115,17 @@ func BuildCompliantName(metric pmetric.Metric, namespace string, addMetricSuffix
 	return metricName
 }
 
-// Build a normalized name for the specified metric
+// Build a normalized name for the specified metric.
 func normalizeName(metric pmetric.Metric, namespace string) string {
-	// Split metric name into "tokens" (remove all non-alphanumerics)
+	// Regexp for characters that can't be in a metric name token.
+	nonTokenMetricCharRE := regexp.MustCompile(`[^a-zA-Z0-9:]`)
+
+	// Split metric name into "tokens" (of supported metric name runes).
+	// Note that this has the side effect of replacing multiple consecutive underscores with a single underscore.
+	// This is part of the OTel to Prometheus specification: https://github.com/open-telemetry/opentelemetry-specification/blob/v1.38.0/specification/compatibility/prometheus_and_openmetrics.md#otlp-metric-points-to-prometheus.
 	nameTokens := strings.FieldsFunc(
 		metric.Name(),
-		func(r rune) bool { return !unicode.IsLetter(r) && !unicode.IsDigit(r) && r != ':' },
+		func(r rune) bool { return nonTokenMetricCharRE.MatchString(string(r)) },
 	)
 
 	// Split unit at the '/' if any
@@ -123,11 +134,12 @@ func normalizeName(metric pmetric.Metric, namespace string) string {
 	// Main unit
 	// Append if not blank, doesn't contain '{}', and is not present in metric name already
 	if len(unitTokens) > 0 {
+		var mainUnitProm, perUnitProm string
 		mainUnitOTel := strings.TrimSpace(unitTokens[0])
 		if mainUnitOTel != "" && !strings.ContainsAny(mainUnitOTel, "{}") {
-			mainUnitProm := CleanUpString(unitMapGetOrDefault(mainUnitOTel))
-			if mainUnitProm != "" && !contains(nameTokens, mainUnitProm) {
-				nameTokens = append(nameTokens, mainUnitProm)
+			mainUnitProm = cleanUpUnit(unitMapGetOrDefault(mainUnitOTel))
+			if slices.Contains(nameTokens, mainUnitProm) {
+				mainUnitProm = ""
 			}
 		}
 
@@ -136,13 +148,26 @@ func normalizeName(metric pmetric.Metric, namespace string) string {
 		if len(unitTokens) > 1 && unitTokens[1] != "" {
 			perUnitOTel := strings.TrimSpace(unitTokens[1])
 			if perUnitOTel != "" && !strings.ContainsAny(perUnitOTel, "{}") {
-				perUnitProm := CleanUpString(perUnitMapGetOrDefault(perUnitOTel))
-				if perUnitProm != "" && !contains(nameTokens, perUnitProm) {
-					nameTokens = append(nameTokens, "per", perUnitProm)
+				perUnitProm = cleanUpUnit(perUnitMapGetOrDefault(perUnitOTel))
+			}
+			if perUnitProm != "" {
+				perUnitProm = "per_" + perUnitProm
+				if slices.Contains(nameTokens, perUnitProm) {
+					perUnitProm = ""
 				}
 			}
 		}
 
+		if perUnitProm != "" {
+			mainUnitProm = strings.TrimSuffix(mainUnitProm, "_")
+		}
+
+		if mainUnitProm != "" {
+			nameTokens = append(nameTokens, mainUnitProm)
+		}
+		if perUnitProm != "" {
+			nameTokens = append(nameTokens, perUnitProm)
+		}
 	}
 
 	// Append _total for Counters
@@ -235,15 +260,15 @@ func removeSuffix(tokens []string, suffix string) []string {
 	return tokens
 }
 
-// Clean up specified string so it's Prometheus compliant
-func CleanUpString(s string) string {
-	return strings.Join(strings.FieldsFunc(s, func(r rune) bool { return !unicode.IsUpper(r) && !unicode.IsLower(r) && !unicode.IsDigit(r) }), "_")
-}
-
-func RemovePromForbiddenRunes(s string) string {
-	return strings.Join(strings.FieldsFunc(s, func(r rune) bool {
-		return !unicode.IsUpper(r) && !unicode.IsLower(r) && !unicode.IsDigit(r) && r != '_' && r != ':'
-	}), "_")
+// cleanUpUnit cleans up unit so it matches model.LabelNameRE.
+func cleanUpUnit(unit string) string {
+	// Multiple consecutive underscores are replaced with a single underscore.
+	// This is part of the OTel to Prometheus specification: https://github.com/open-telemetry/opentelemetry-specification/blob/v1.38.0/specification/compatibility/prometheus_and_openmetrics.md#otlp-metric-points-to-prometheus.
+	multipleUnderscoresRE := regexp.MustCompile(`__+`)
+	return strings.TrimPrefix(multipleUnderscoresRE.ReplaceAllString(
+		strutil.SanitizeLabelName(unit),
+		"_",
+	), "_")
 }
 
 // Retrieve the Prometheus "basic" unit corresponding to the specified "basic" unit
@@ -264,16 +289,6 @@ func perUnitMapGetOrDefault(perUnit string) string {
 	return perUnit
 }
 
-// Returns whether the slice contains the specified value
-func contains(slice []string, value string) bool {
-	for _, sliceEntry := range slice {
-		if sliceEntry == value {
-			return true
-		}
-	}
-	return false
-}
-
 // Remove the specified value from the slice
 func removeItem(slice []string, value string) []string {
 	newSlice := make([]string, 0, len(slice))

storage/remote/otlptranslator/prometheus/normalize_name_test.go

@@ -148,13 +148,13 @@ func TestNamespace(t *testing.T) {
 	require.Equal(t, "space_test", normalizeName(createGauge("#test", ""), "space"))
 }
 
-func TestCleanUpString(t *testing.T) {
-	require.Equal(t, "", CleanUpString(""))
-	require.Equal(t, "a_b", CleanUpString("a b"))
-	require.Equal(t, "hello_world", CleanUpString("hello, world!"))
-	require.Equal(t, "hello_you_2", CleanUpString("hello you 2"))
-	require.Equal(t, "1000", CleanUpString("$1000"))
-	require.Equal(t, "", CleanUpString("*+$^=)"))
+func TestCleanUpUnit(t *testing.T) {
+	require.Equal(t, "", cleanUpUnit(""))
+	require.Equal(t, "a_b", cleanUpUnit("a b"))
+	require.Equal(t, "hello_world", cleanUpUnit("hello, world"))
+	require.Equal(t, "hello_you_2", cleanUpUnit("hello you 2"))
+	require.Equal(t, "1000", cleanUpUnit("$1000"))
+	require.Equal(t, "", cleanUpUnit("*+$^=)"))
 }
 
 func TestUnitMapGetOrDefault(t *testing.T) {
@@ -179,7 +179,7 @@ func TestRemoveItem(t *testing.T) {
 	require.Equal(t, []string{"b", "c"}, removeItem([]string{"a", "b", "c"}, "a"))
 }
 
-func TestBuildCompliantNameWithNormalize(t *testing.T) {
+func TestBuildCompliantNameWithSuffixes(t *testing.T) {
 	require.Equal(t, "system_io_bytes_total", BuildCompliantName(createCounter("system.io", "By"), "", true))
 	require.Equal(t, "system_network_io_bytes_total", BuildCompliantName(createCounter("network.io", "By"), "system", true))
 	require.Equal(t, "_3_14_digits", BuildCompliantName(createGauge("3.14 digits", ""), "", true))
@@ -190,6 +190,7 @@ func TestBuildCompliantNameWithNormalize(t *testing.T) {
 	require.Equal(t, "foo_bar_ratio", BuildCompliantName(createGauge("foo.bar", "1"), "", true))
 	// Slashes in units are converted.
 	require.Equal(t, "system_io_foo_per_bar_total", BuildCompliantName(createCounter("system.io", "foo/bar"), "", true))
+	require.Equal(t, "metric_with_foreign_characters_total", BuildCompliantName(createCounter("metric_with_字符_foreign_characters", ""), "", true))
 }
 
 func TestBuildCompliantNameWithoutSuffixes(t *testing.T) {

storage/remote/otlptranslator/prometheusremotewrite/helper_test.go

@@ -48,7 +48,6 @@ func TestCreateAttributes(t *testing.T) {
 		resource.Attributes().PutStr(k, v)
 	}
 	attrs := pcommon.NewMap()
-	attrs.PutStr("__name__", "test_metric")
 	attrs.PutStr("metric-attr", "metric value")
 
 	testCases := []struct {
@@ -162,7 +161,7 @@ func TestCreateAttributes(t *testing.T) {
 			settings := Settings{
 				PromoteResourceAttributes: tc.promoteResourceAttributes,
 			}
-			lbls := createAttributes(resource, attrs, settings, nil, false)
+			lbls := createAttributes(resource, attrs, settings, nil, false, model.MetricNameLabel, "test_metric")
 
 			assert.ElementsMatch(t, lbls, tc.expectedLabels)
 		})