Using google sdk to discover the project ID if it was not provided (#62)

* moving back the project ID to the main package, as, with the auto discovery, we want that to be defined at boottime, blocking the startup if needed, rather than doing it on each collector creation / handler. Simple project-id discovery logic using the gcloud sdk

Signed-off-by: Daniel Caballero <[email protected]>

Author: dcaba

README.md

@@ -63,7 +63,7 @@ If you are still using the legacy [Access scopes][access-scopes], the `https://w
 
 | Flag / Environment Variable | Required | Default | Description |
 | --------------------------- | -------- | ------- | ----------- |
-| `google.project-id`<br />`STACKDRIVER_EXPORTER_GOOGLE_PROJECT_ID` | Yes | | Google Project ID |
+| `google.project-id`<br />`STACKDRIVER_EXPORTER_GOOGLE_PROJECT_ID` | No | GCloud SDK autodiscovery | Google Project ID |
 | `monitoring.metrics-type-prefixes`<br />`STACKDRIVER_EXPORTER_MONITORING_METRICS_TYPE_PREFIXES` | Yes | | Comma separated Google Stackdriver Monitoring Metric Type prefixes (see [example][metrics-prefix-example] and [available metrics][metrics-list]) |
 | `monitoring.metrics-interval`<br />`STACKDRIVER_EXPORTER_MONITORING_METRICS_INTERVAL` | No | `5m` | Metric's timestamp interval to request from the Google Stackdriver Monitoring Metrics API. Only the most recent data point is used |
 | `monitoring.metrics-offset`<br />`STACKDRIVER_EXPORTER_MONITORING_METRICS_OFFSET` | No | `0s` | Offset (into the past) for the metric's timestamp interval to request from the Google Stackdriver Monitoring Metrics API, to handle latency in published metrics |

collectors/monitoring_collector.go

@@ -33,10 +33,6 @@ import (
 )
 
 var (
-	projectID = kingpin.Flag(
-		"google.project-id", "Google Project ID ($STACKDRIVER_EXPORTER_GOOGLE_PROJECT_ID).",
-	).Envar("STACKDRIVER_EXPORTER_GOOGLE_PROJECT_ID").Required().String()
-
 	monitoringMetricsTypePrefixes = kingpin.Flag(
 		"monitoring.metrics-type-prefixes", "Comma separated Google Stackdriver Monitoring Metric Type prefixes ($STACKDRIVER_EXPORTER_MONITORING_METRICS_TYPE_PREFIXES).",
 	).Envar("STACKDRIVER_EXPORTER_MONITORING_METRICS_TYPE_PREFIXES").Required().String()
@@ -75,11 +71,7 @@ type MonitoringCollector struct {
 	logger                          log.Logger
 }
 
-func NewMonitoringCollector(monitoringService *monitoring.Service, filters map[string]bool, logger log.Logger) (*MonitoringCollector, error) {
-	if *projectID == "" {
-		return nil, errors.New("Flag `google.project-id` is required")
-	}
-
+func NewMonitoringCollector(projectID string, monitoringService *monitoring.Service, filters map[string]bool, logger log.Logger) (*MonitoringCollector, error) {
 	if *monitoringMetricsTypePrefixes == "" {
 		return nil, errors.New("Flag `monitoring.metrics-type-prefixes` is required")
 	}
@@ -90,7 +82,7 @@ func NewMonitoringCollector(monitoringService *monitoring.Service, filters map[s
 			Subsystem:   "monitoring",
 			Name:        "api_calls_total",
 			Help:        "Total number of Google Stackdriver Monitoring API calls made.",
-			ConstLabels: prometheus.Labels{"project_id": *projectID},
+			ConstLabels: prometheus.Labels{"project_id": projectID},
 		},
 	)
 
@@ -100,7 +92,7 @@ func NewMonitoringCollector(monitoringService *monitoring.Service, filters map[s
 			Subsystem:   "monitoring",
 			Name:        "scrapes_total",
 			Help:        "Total number of Google Stackdriver Monitoring metrics scrapes.",
-			ConstLabels: prometheus.Labels{"project_id": *projectID},
+			ConstLabels: prometheus.Labels{"project_id": projectID},
 		},
 	)
 
@@ -110,7 +102,7 @@ func NewMonitoringCollector(monitoringService *monitoring.Service, filters map[s
 			Subsystem:   "monitoring",
 			Name:        "scrape_errors_total",
 			Help:        "Total number of Google Stackdriver Monitoring metrics scrape errors.",
-			ConstLabels: prometheus.Labels{"project_id": *projectID},
+			ConstLabels: prometheus.Labels{"project_id": projectID},
 		},
 	)
 
@@ -120,7 +112,7 @@ func NewMonitoringCollector(monitoringService *monitoring.Service, filters map[s
 			Subsystem:   "monitoring",
 			Name:        "last_scrape_error",
 			Help:        "Whether the last metrics scrape from Google Stackdriver Monitoring resulted in an error (1 for error, 0 for success).",
-			ConstLabels: prometheus.Labels{"project_id": *projectID},
+			ConstLabels: prometheus.Labels{"project_id": projectID},
 		},
 	)
 
@@ -130,7 +122,7 @@ func NewMonitoringCollector(monitoringService *monitoring.Service, filters map[s
 			Subsystem:   "monitoring",
 			Name:        "last_scrape_timestamp",
 			Help:        "Number of seconds since 1970 since last metrics scrape from Google Stackdriver Monitoring.",
-			ConstLabels: prometheus.Labels{"project_id": *projectID},
+			ConstLabels: prometheus.Labels{"project_id": projectID},
 		},
 	)
 
@@ -140,7 +132,7 @@ func NewMonitoringCollector(monitoringService *monitoring.Service, filters map[s
 			Subsystem:   "monitoring",
 			Name:        "last_scrape_duration_seconds",
 			Help:        "Duration of the last metrics scrape from Google Stackdriver Monitoring.",
-			ConstLabels: prometheus.Labels{"project_id": *projectID},
+			ConstLabels: prometheus.Labels{"project_id": projectID},
 		},
 	)
 
@@ -156,7 +148,7 @@ func NewMonitoringCollector(monitoringService *monitoring.Service, filters map[s
 	}
 
 	monitoringCollector := &MonitoringCollector{
-		projectID:                       *projectID,
+		projectID:                       projectID,
 		metricsTypePrefixes:             filteredPrefixes,
 		metricsInterval:                 *monitoringMetricsInterval,
 		metricsOffset:                   *monitoringMetricsOffset,

stackdriver_exporter.go

@@ -28,6 +28,7 @@ import (
 	"github.com/prometheus/common/version"
 	"golang.org/x/net/context"
 	"golang.org/x/oauth2/google"
+	"google.golang.org/api/compute/v1"
 	"google.golang.org/api/monitoring/v3"
 	"google.golang.org/api/option"
 	"gopkg.in/alecthomas/kingpin.v2"
@@ -44,6 +45,10 @@ var (
 		"web.telemetry-path", "Path under which to expose Prometheus metrics ($STACKDRIVER_EXPORTER_WEB_TELEMETRY_PATH).",
 	).Envar("STACKDRIVER_EXPORTER_WEB_TELEMETRY_PATH").Default("/metrics").String()
 
+	projectID = kingpin.Flag(
+		"google.project-id", "Google Project ID ($STACKDRIVER_EXPORTER_GOOGLE_PROJECT_ID).",
+	).Envar("STACKDRIVER_EXPORTER_GOOGLE_PROJECT_ID").String()
+
 	stackdriverMaxRetries = kingpin.Flag(
 		"stackdriver.max-retries", "Max number of retries that should be attempted on 503 errors from stackdriver. ($STACKDRIVER_EXPORTER_MAX_RETRIES)",
 	).Envar("STACKDRIVER_EXPORTER_MAX_RETRIES").Default("0").Int()
@@ -69,9 +74,18 @@ func init() {
 	prometheus.MustRegister(version.NewCollector("stackdriver_exporter"))
 }
 
-func createMonitoringService() (*monitoring.Service, error) {
-	ctx := context.Background()
+func getDefaultGCPProject(ctx context.Context) (*string, error) {
+	credentials, err := google.FindDefaultCredentials(ctx, compute.ComputeScope)
+	if err != nil {
+		return nil, err
+	}
+	if credentials.ProjectID == "" {
+		return nil, fmt.Errorf("unable to identify the gcloud project. Got empty string")
+	}
+	return &credentials.ProjectID, nil
+}
 
+func createMonitoringService(ctx context.Context) (*monitoring.Service, error) {
 	googleClient, err := google.DefaultClient(ctx, monitoring.MonitoringReadScope)
 	if err != nil {
 		return nil, fmt.Errorf("Error creating Google client: %v", err)
@@ -94,7 +108,7 @@ func createMonitoringService() (*monitoring.Service, error) {
 	return monitoringService, nil
 }
 
-func newHandler(m *monitoring.Service, logger log.Logger) http.HandlerFunc {
+func newHandler(projectID string, m *monitoring.Service, logger log.Logger) http.HandlerFunc {
 	return func(w http.ResponseWriter, r *http.Request) {
 		collectParams := r.URL.Query()["collect"]
 
@@ -104,7 +118,7 @@ func newHandler(m *monitoring.Service, logger log.Logger) http.HandlerFunc {
 			filters[param] = true
 		}
 
-		monitoringCollector, err := collectors.NewMonitoringCollector(m, filters, logger)
+		monitoringCollector, err := collectors.NewMonitoringCollector(projectID, m, filters, logger)
 		if err != nil {
 			level.Error(logger).Log("err", err)
 			os.Exit(1)
@@ -133,16 +147,28 @@ func main() {
 
 	logger := promlog.New(promlogConfig)
 
+	ctx := context.Background()
+	if *projectID == "" {
+		level.Info(logger).Log("msg", "no projectID was provided. Trying to discover it")
+		var err error
+		projectID, err = getDefaultGCPProject(ctx)
+		if err != nil {
+			level.Error(logger).Log("msg", "no explicit projectID and error trying to discover default GCloud project", "err", err)
+			os.Exit(1)
+		}
+	}
+
 	level.Info(logger).Log("msg", "Starting stackdriver_exporter", "version", version.Info())
 	level.Info(logger).Log("msg", "Build context", "build_context", version.BuildContext())
+	level.Info(logger).Log("msg", "Using Google Cloud Project ID", "projectID", *projectID)
 
-	monitoringService, err := createMonitoringService()
+	monitoringService, err := createMonitoringService(ctx)
 	if err != nil {
 		level.Error(logger).Log("msg", "failed to create monitoring service", "err", err)
 		os.Exit(1)
 	}
 
-	handlerFunc := newHandler(monitoringService, logger)
+	handlerFunc := newHandler(*projectID, monitoringService, logger)
 
 	http.Handle(*metricsPath, promhttp.InstrumentMetricHandler(prometheus.DefaultRegisterer, handlerFunc))
 	http.HandleFunc("/", func(w http.ResponseWriter, r *http.Request) {