Change delta calcaulator defaults (#52)

Changes the values for the telemetry emitter delta calculator expiration
age and expiration check interval from 30 seconds to 5 minutes. Also
adds the following configuration options for setting each setting:

- telemetry_emitter_delta_expiration_age
- telemetry_emitter_delta_expiration_check_interval

This fixes the issue of missing datapoints for counters when the scrape
interval was taking longer than the delta expiration check interval and age.

CHANGELOG.md

@@ -5,6 +5,17 @@ All notable changes to this project will be documented in this file.
 The format is based on [Keep a Changelog](http://keepachangelog.com/)
 and this project adheres to [Semantic Versioning](http://semver.org/).
 
+## NEXT RELEASE
+### Changed
+- Change the default for the New Relic telemetry emitter delta calculator 
+  expiration age and expiration check interval to 5 minutes. These values can 
+  now be configured with the following options
+  `telemetry_emitter_delta_expiration_age` and
+  `telemetry_emitter_delta_expiration_check_interval`. This solves the issue
+  with missing counters when the scrape interval takes more than the hard-coded
+  30 minutes values that we had previously.
+- Use go modules instead of govendor for managing project dependencies.
+
 ## 1.4.0
 ### Changed
 - Redact the password from URLs using simple authentication when storing target metadata.

deploy/nri-prometheus.tmpl.yaml

@@ -89,6 +89,14 @@ data:
     # The HTTP client timeout when fetching data from endpoints. Defaults to 5s.
     # scrape_timeout: "5s"
 
+    # How old must the entries used for calculating the counters delta be
+    # before the telemetry emitter expires them. Defaults to 5m.
+    # telemetry_emitter_delta_expiration_age: "5m"
+
+    # How often must the telemetry emitter check for expired delta entries.
+    # Defaults to 5m.
+    # telemetry_emitter_delta_expiration_check_interval: "5m"
+
     # Wether the integration should run in verbose mode or not. Defaults to false.
     verbose: false
 

internal/cmd/scraper/scraper.go

@@ -44,9 +44,11 @@ type Config struct {
 	DecorateFile                      bool
 	EmitterProxy                      string `mapstructure:"emitter_proxy"`
 	// Parsed version of `EmitterProxy`
-	EmitterProxyURL           *url.URL
-	EmitterCAFile             string `mapstructure:"emitter_ca_file"`
-	EmitterInsecureSkipVerify bool   `mapstructure:"emitter_insecure_skip_verify" default:"false"`
+	EmitterProxyURL                              *url.URL
+	EmitterCAFile                                string        `mapstructure:"emitter_ca_file"`
+	EmitterInsecureSkipVerify                    bool          `mapstructure:"emitter_insecure_skip_verify" default:"false"`
+	TelemetryEmitterDeltaExpirationAge           time.Duration `mapstructure:"telemetry_emitter_delta_expiration_age"`
+	TelemetryEmitterDeltaExpirationCheckInterval time.Duration `mapstructure:"telemetry_emitter_delta_expiration_check_interval"`
 }
 
 const maskedLicenseKey = "****"
@@ -108,9 +110,6 @@ func validateConfig(cfg *Config) error {
 // RunWithEmitters runs the scraper with preselected emitters.
 func RunWithEmitters(cfg *Config, emitters []integration.Emitter) error {
 	logrus.Infof("Starting New Relic's Prometheus OpenMetrics Integration version %s", integration.Version)
-	if cfg.Verbose {
-		logrus.SetLevel(logrus.DebugLevel)
-	}
 	logrus.Debugf("Config: %#v", cfg)
 
 	if len(emitters) == 0 {
@@ -185,6 +184,9 @@ func Run(cfg *Config) error {
 	if err != nil {
 		return fmt.Errorf("while getting configuration options: %w", err)
 	}
+	if cfg.Verbose {
+		logrus.SetLevel(logrus.DebugLevel)
+	}
 
 	var emitters []integration.Emitter
 	for _, e := range cfg.Emitters {
@@ -242,8 +244,10 @@ func Run(cfg *Config) error {
 			}
 
 			c := integration.TelemetryEmitterConfig{
-				Percentiles:   cfg.Percentiles,
-				HarvesterOpts: harvesterOpts,
+				Percentiles:                   cfg.Percentiles,
+				HarvesterOpts:                 harvesterOpts,
+				DeltaExpirationAge:            cfg.TelemetryEmitterDeltaExpirationAge,
+				DeltaExpirationCheckInternval: cfg.TelemetryEmitterDeltaExpirationCheckInterval,
 			}
 
 			emitter, err := integration.NewTelemetryEmitter(c)

internal/integration/emitter.go

@@ -21,9 +21,8 @@ import (
 )
 
 const (
-	// Ideally these 2 values should be configurable as integer multiples of the scrape interval.
-	defaultDeltaExpirationAge           = 30 * time.Second
-	defaultDeltaExpirationCheckInterval = 30 * time.Second
+	defaultDeltaExpirationAge           = 5 * time.Minute
+	defaultDeltaExpirationCheckInterval = 5 * time.Minute
 )
 
 // Emitter is an interface representing the ability to emit metrics.
@@ -147,24 +146,24 @@ func TelemetryHarvesterWithProxy(proxyURL *url.URL) TelemetryHarvesterOpt {
 func NewTelemetryEmitter(cfg TelemetryEmitterConfig) (*TelemetryEmitter, error) {
 	dc := cumulative.NewDeltaCalculator()
 
+	deltaExpirationAge := defaultDeltaExpirationAge
 	if cfg.DeltaExpirationAge != 0 {
-		dc.SetExpirationAge(cfg.DeltaExpirationAge)
-	} else {
-		dc.SetExpirationAge(defaultDeltaExpirationAge)
+		deltaExpirationAge = cfg.DeltaExpirationAge
 	}
+	dc.SetExpirationAge(deltaExpirationAge)
 	logrus.Debugf(
 		"telemetry emitter configured with delta counter expiration age: %s",
-		cfg.DeltaExpirationAge,
+		deltaExpirationAge,
 	)
 
+	deltaExpirationCheckInterval := defaultDeltaExpirationCheckInterval
 	if cfg.DeltaExpirationCheckInternval != 0 {
-		dc.SetExpirationCheckInterval(cfg.DeltaExpirationCheckInternval)
-	} else {
-		dc.SetExpirationCheckInterval(defaultDeltaExpirationCheckInterval)
+		deltaExpirationCheckInterval = cfg.DeltaExpirationCheckInternval
 	}
+	dc.SetExpirationCheckInterval(deltaExpirationCheckInterval)
 	logrus.Debugf(
 		"telemetry emitter configured with delta counter expiration check interval: %s",
-		cfg.DeltaExpirationAge,
+		deltaExpirationCheckInterval,
 	)
 
 	harvester, err := telemetry.NewHarvester(cfg.HarvesterOpts...)