scrape: Add config option for escaping scheme request. (prometheus#16066)

The new metric_name_escaping_scheme config option works in parallel with metric_name_validation_scheme and controls which escaping scheme is requested when scraping. When not specified, the scheme will request underscores if the validation scheme is set to legacy, and will request allow-utf-8 when the validation scheme is set to utf8. This setting allows users to allow utf8 names if they like, but explicitly request an escaping scheme rather than UTF-8.

Fixes prometheus#16034

Built on prometheus#16080

Signed-off-by: Owen Williams <owen.williams@grafana.com>

Author: ywwg

config/config.go

@@ -479,8 +479,13 @@ type GlobalConfig struct {
 	// Keep no more than this many dropped targets per job.
 	// 0 means no limit.
 	KeepDroppedTargets uint `yaml:"keep_dropped_targets,omitempty"`
-	// Allow UTF8 Metric and Label Names.
+	// Allow UTF8 Metric and Label Names. Can be blank in config files but must
+	// have a value if a ScrepeConfig is created programmatically.
 	MetricNameValidationScheme string `yaml:"metric_name_validation_scheme,omitempty"`
+	// Metric name escaping mode to request through content negotiation. Can be
+	// blank in config files but must have a value if a ScrepeConfig is created
+	// programmatically.
+	MetricNameEscapingScheme string `yaml:"metric_name_escaping_scheme,omitempty"`
 }
 
 // ScrapeProtocol represents supported protocol for scraping metrics.
@@ -719,8 +724,13 @@ type ScrapeConfig struct {
 	// Keep no more than this many dropped targets per job.
 	// 0 means no limit.
 	KeepDroppedTargets uint `yaml:"keep_dropped_targets,omitempty"`
-	// Allow UTF8 Metric and Label Names.
+	// Allow UTF8 Metric and Label Names. Can be blank in config files but must
+	// have a value if a ScrepeConfig is created programmatically.
 	MetricNameValidationScheme string `yaml:"metric_name_validation_scheme,omitempty"`
+	// Metric name escaping mode to request through content negotiation. Can be
+	// blank in config files but must have a value if a ScrepeConfig is created
+	// programmatically.
+	MetricNameEscapingScheme string `yaml:"metric_name_escaping_scheme,omitempty"`
 
 	// We cannot do proper Go type embedding below as the parser will then parse
 	// values arbitrarily into the overflow maps of further-down types.
@@ -841,13 +851,48 @@ func (c *ScrapeConfig) Validate(globalConfig GlobalConfig) error {
 	if model.NameValidationScheme != model.UTF8Validation {
 		return errors.New("model.NameValidationScheme must be set to UTF8")
 	}
+
 	switch globalConfig.MetricNameValidationScheme {
-	case "", LegacyValidationConfig, UTF8ValidationConfig:
+	case "":
+		globalConfig.MetricNameValidationScheme = UTF8ValidationConfig
+	case LegacyValidationConfig, UTF8ValidationConfig:
 	default:
-		return fmt.Errorf("unknown name validation method specified, must be either 'legacy' or 'utf8', got %s", globalConfig.MetricNameValidationScheme)
+		return fmt.Errorf("unknown global name validation method specified, must be either 'legacy' or 'utf8', got %s", globalConfig.MetricNameValidationScheme)
 	}
-	if c.MetricNameValidationScheme == "" {
+	// Scrapeconfig validation scheme matches global if left blank.
+	switch c.MetricNameValidationScheme {
+	case "":
 		c.MetricNameValidationScheme = globalConfig.MetricNameValidationScheme
+	case LegacyValidationConfig, UTF8ValidationConfig:
+	default:
+		return fmt.Errorf("unknown scrape config name validation method specified, must be either 'legacy' or 'utf8', got %s", c.MetricNameValidationScheme)
+	}
+
+	// Escaping scheme is based on the validation scheme if left blank.
+	switch globalConfig.MetricNameEscapingScheme {
+	case "":
+		if globalConfig.MetricNameValidationScheme == LegacyValidationConfig {
+			globalConfig.MetricNameEscapingScheme = model.EscapeUnderscores
+		} else {
+			globalConfig.MetricNameEscapingScheme = model.AllowUTF8
+		}
+	case model.AllowUTF8, model.EscapeUnderscores, model.EscapeDots, model.EscapeValues:
+	default:
+		return fmt.Errorf("unknown global name escaping method specified, must be one of '%s', '%s', '%s', or '%s', got %s", model.AllowUTF8, model.EscapeUnderscores, model.EscapeDots, model.EscapeValues, globalConfig.MetricNameValidationScheme)
+	}
+
+	if c.MetricNameEscapingScheme == "" {
+		c.MetricNameEscapingScheme = globalConfig.MetricNameEscapingScheme
+	}
+
+	switch c.MetricNameEscapingScheme {
+	case model.AllowUTF8:
+		if c.MetricNameValidationScheme != UTF8ValidationConfig {
+			return errors.New("utf8 metric names requested but validation scheme is not set to UTF8")
+		}
+	case model.EscapeUnderscores, model.EscapeDots, model.EscapeValues:
+	default:
+		return fmt.Errorf("unknown scrape config name escaping method specified, must be one of '%s', '%s', '%s', or '%s', got %s", model.AllowUTF8, model.EscapeUnderscores, model.EscapeDots, model.EscapeValues, c.MetricNameValidationScheme)
 	}
 
 	return nil
@@ -858,6 +903,20 @@ func (c *ScrapeConfig) MarshalYAML() (interface{}, error) {
 	return discovery.MarshalYAMLWithInlineConfigs(c)
 }
 
+// ToValidationScheme returns the validation scheme for the given string config value.
+func ToValidationScheme(s string) (validationScheme model.ValidationScheme, err error) {
+	switch s {
+	case UTF8ValidationConfig:
+		validationScheme = model.UTF8Validation
+	case LegacyValidationConfig:
+		validationScheme = model.LegacyValidation
+	default:
+		return model.UTF8Validation, fmt.Errorf("invalid metric name validation scheme, %s", s)
+	}
+
+	return validationScheme, nil
+}
+
 // StorageConfig configures runtime reloadable configuration options.
 type StorageConfig struct {
 	TSDBConfig      *TSDBConfig      `yaml:"tsdb,omitempty"`