add validation for metrics url

Signed-off-by: ChrsMark <[email protected]>

Author: ChrsMark

cmd/mdatagen/internal/metadata.go

@@ -122,8 +122,8 @@ func (md *Metadata) validateMetricsAndEvents() error {
 	var errs error
 	usedAttrs := map[AttributeName]bool{}
 	errs = errors.Join(errs,
-		validateMetrics(md.Metrics, md.Attributes, usedAttrs),
-		validateMetrics(md.Telemetry.Metrics, md.Attributes, usedAttrs),
+		validateMetrics(md.Metrics, md.Attributes, usedAttrs, md.SemConvVersion),
+		validateMetrics(md.Telemetry.Metrics, md.Attributes, usedAttrs, md.SemConvVersion),
 		validateEvents(md.Events, md.Attributes, usedAttrs),
 		md.validateAttributes(usedAttrs))
 	return errs
@@ -166,10 +166,10 @@ func (md *Metadata) supportsSignal(signal string) bool {
 	return false
 }
 
-func validateMetrics(metrics map[MetricName]Metric, attributes map[AttributeName]Attribute, usedAttrs map[AttributeName]bool) error {
+func validateMetrics(metrics map[MetricName]Metric, attributes map[AttributeName]Attribute, usedAttrs map[AttributeName]bool, semConvVersion string) error {
 	var errs error
 	for mn, m := range metrics {
-		if err := m.validate(); err != nil {
+		if err := m.validate(mn, semConvVersion); err != nil {
 			errs = errors.Join(errs, fmt.Errorf(`metric "%v": %w`, mn, err))
 			continue
 		}

cmd/mdatagen/internal/metric.go

@@ -6,6 +6,7 @@ package internal // import "go.opentelemetry.io/collector/cmd/mdatagen/internal"
 import (
 	"errors"
 	"fmt"
+	"regexp"
 	"strings"
 
 	"golang.org/x/text/cases"
@@ -16,6 +17,8 @@ import (
 	"go.opentelemetry.io/collector/pdata/pmetric"
 )
 
+var reNonAlnum = regexp.MustCompile(`[^a-z0-9]+`)
+
 type MetricName string
 
 func (mn MetricName) Render() (string, error) {
@@ -62,7 +65,7 @@ func (s Stability) String() string {
 	return fmt.Sprintf(" [%s]", s.Level)
 }
 
-func (m *Metric) validate() error {
+func (m *Metric) validate(metricName MetricName, semConvVersion string) error {
 	var errs error
 	if m.Sum == nil && m.Gauge == nil && m.Histogram == nil {
 		errs = errors.Join(errs, errors.New("missing metric type key, "+
@@ -84,9 +87,48 @@ func (m *Metric) validate() error {
 	if m.Gauge != nil {
 		errs = errors.Join(errs, m.Gauge.Validate())
 	}
+	if m.SemanticConvention != nil {
+		if err := validateSemConvMetricURL(m.SemanticConvention.SemanticConventionRef, semConvVersion, string(metricName)); err != nil {
+			errs = errors.Join(errs, err)
+		}
+	}
 	return errs
 }
 
+func metricAnchor(metricName string) string {
+	m := strings.ToLower(strings.TrimSpace(metricName))
+	m = reNonAlnum.ReplaceAllString(m, "")
+	return "metric-" + m
+}
+
+// validateSemConvMetricURL verifies the URL matches exactly:
+// https://github.com/open-telemetry/semantic-conventions/blob/<semConvVersion>/*#metric-<metricName>
+func validateSemConvMetricURL(rawURL, semConvVersion, metricName string) error {
+	if strings.TrimSpace(rawURL) == "" {
+		return fmt.Errorf("url is empty")
+	}
+	if strings.TrimSpace(semConvVersion) == "" {
+		return fmt.Errorf("semConvVersion is empty")
+	}
+	if strings.TrimSpace(metricName) == "" {
+		return fmt.Errorf("metricName is empty")
+	}
+	semConvVersion = "v" + semConvVersion
+
+	anchor := metricAnchor(metricName)
+	// Build a strict regex that enforces https, repo, blob, given version, any doc path, and exact anchor.
+	pattern := fmt.Sprintf(`^https://github\.com/open-telemetry/semantic-conventions/blob/%s/[^#\s]+#%s$`,
+		regexp.QuoteMeta(semConvVersion),
+		regexp.QuoteMeta(anchor),
+	)
+	re := regexp.MustCompile(pattern)
+	if !re.MatchString(rawURL) {
+		return fmt.Errorf("invalid semantic-conventions URL: want https://github.com/open-telemetry/semantic-conventions/blob/%s/*#%s, got %q",
+			semConvVersion, anchor, rawURL)
+	}
+	return nil
+}
+
 func (m *Metric) Unmarshal(parser *confmap.Conf) error {
 	if !parser.IsSet("enabled") {
 		return errors.New("missing required field: `enabled`")

cmd/mdatagen/internal/samplescraper/documentation.md

@@ -68,7 +68,7 @@ The metric will be become optional soon.
 
 | Unit | Metric Type | Value Type | Aggregation Temporality | Monotonic | Stability | Semantic Convention |
 | ---- | ----------- | ---------- | ----------------------- | --------- | --------- | ------------------- |
-| s | Sum | Int | Cumulative | true | beta | [system.cpu.time](https://github.com/open-telemetry/semantic-conventions/blob/v1.35.0/docs/system/system-metrics.md#metric-systemcputime) |
+| s | Sum | Int | Cumulative | true | beta | [system.cpu.time](https://github.com/open-telemetry/semantic-conventions/blob/v1.37.0/docs/system/system-metrics.md#metric-systemcputime) |
 
 ## Optional Metrics
 

cmd/mdatagen/internal/samplescraper/internal/metadata/generated_logs.go

@@ -3,7 +3,7 @@
 type: sample
 github_project: open-telemetry/opentelemetry-collector
 
-sem_conv_version: 1.9.0
+sem_conv_version: 1.37.0
 
 status:
   disable_codecov_badge: true
@@ -112,7 +112,7 @@ metrics:
       monotonic: true
       aggregation_temporality: cumulative
     semantic_convention:
-      semconv_ref: https://github.com/open-telemetry/semantic-conventions/blob/v1.35.0/docs/system/system-metrics.md#metric-systemcputime
+      semconv_ref: https://github.com/open-telemetry/semantic-conventions/blob/v1.37.0/docs/system/system-metrics.md#metric-systemcputime
   default.metric:
     enabled: true
     description: Monotonic cumulative sum int metric enabled by default.