Add stackdriver metrics analysis

Signed-off-by: Somtochi Onyekwere <somtochionyekwere@gmail.com>

artifacts/flagger/crd.yaml

@@ -1104,6 +1104,7 @@ spec:
                         - prometheus
                         - influxdb
                         - datadog
+                        - stackdriver
                         - cloudwatch
                         - newrelic
                         - graphite

charts/flagger/crds/crd.yaml

@@ -1104,6 +1104,7 @@ spec:
                         - prometheus
                         - influxdb
                         - datadog
+                        - stackdriver
                         - cloudwatch
                         - newrelic
                         - graphite

docs/gitbook/usage/metrics.md

@@ -477,3 +477,53 @@ spec:
     secretRef:
       name: graphite-basic-auth
 ```
+
+## Google CLoud Monitoring (Stackdriver)
+
+Enable Workload Identity on your cluster, create a service account key that has read access to the
+Cloud Monitoring API and then create an IAM policy binding between the GCP service account and the Flagger 
+service account on Kubernetes. You can take a look at this [guide](https://cloud.google.com/kubernetes-engine/docs/how-to/workload-identity)
+
+Annotate the flagger service account
+```shell script
+kubectl annotate serviceaccount flagger \
+    --namespace <namespace> \
+    iam.gke.io/gcp-service-account=<gcp-serviceaccount-name>@<project-id>.iam.gserviceaccount.com
+```
+
+Alternatively, you can download the json keys and add it to your secret with the key `serviceAccountKey` (This method is not recommended).
+
+Create a secret that contains your project-id (and, if workload identity is not enabled on your cluster,
+your [service account json](https://cloud.google.com/docs/authentication/production#create_service_account)).
+
+```
+ kubectl create secret generic gcloud-sa --from-literal=project=<project-id>
+```
+
+Then reference the secret in the metric template. 
+Note: The particular MQL query used here works if [Istio is installed on GKE](https://cloud.google.com/istio/docs/istio-on-gke/installing).
+```yaml
+apiVersion: flagger.app/v1beta1
+kind: MetricTemplate
+metadata:
+  name: bytes-sent
+  namespace: test
+spec:
+  provider:
+    type: stackdriver
+    secretRef: 
+      name: gcloud-sa
+  query: |
+    fetch k8s_container
+    | metric 'istio.io/service/server/response_latencies'
+    | filter
+        (metric.destination_service_name == '{{ service }}-canary'
+        && metric.destination_service_namespace == '{{ namespace }}')
+    | align delta(1m)
+    | every 1m
+    | group_by [],
+        [value_response_latencies_percentile:
+          percentile(value.response_latencies, 99)]
+```
+
+The reference for the query language can be found [here](https://cloud.google.com/monitoring/mql/reference)

go.mod

@@ -3,14 +3,20 @@ module github.com/fluxcd/flagger
 go 1.16
 
 require (
+	cloud.google.com/go/monitoring v0.1.0
 	github.com/Masterminds/semver/v3 v3.0.3
 	github.com/aws/aws-sdk-go v1.37.32
 	github.com/davecgh/go-spew v1.1.1
 	github.com/go-logr/zapr v0.3.0
-	github.com/google/go-cmp v0.5.5
+	github.com/google/go-cmp v0.5.6
+	github.com/googleapis/gax-go/v2 v2.0.5
 	github.com/prometheus/client_golang v1.11.0
 	github.com/stretchr/testify v1.7.0
 	go.uber.org/zap v1.14.1
+	google.golang.org/api v0.54.0
+	google.golang.org/genproto v0.0.0-20210813162853-db860fec028c
+	google.golang.org/grpc v1.39.1
+	google.golang.org/protobuf v1.27.1
 	gopkg.in/h2non/gock.v1 v1.0.15
 	k8s.io/api v0.21.1
 	k8s.io/apimachinery v0.21.1

kustomize/base/flagger/crd.yaml

@@ -1104,6 +1104,7 @@ spec:
                         - prometheus
                         - influxdb
                         - datadog
+                        - stackdriver
                         - cloudwatch
                         - newrelic
                         - graphite

pkg/metrics/providers/factory.go

@@ -38,6 +38,8 @@ func (factory Factory) Provider(
 		return NewNewRelicProvider(metricInterval, provider, credentials)
 	case "graphite":
 		return NewGraphiteProvider(provider, credentials)
+	case "stackdriver":
+		return NewStackDriverProvider(provider, credentials)
 	default:
 		return NewPrometheusProvider(provider, credentials)
 	}

pkg/metrics/providers/stackdriver.go

@@ -0,0 +1,129 @@
+package providers
+
+import (
+	"context"
+	"fmt"
+
+	monitoring "cloud.google.com/go/monitoring/apiv3/v2"
+	"google.golang.org/api/iterator"
+	"google.golang.org/api/option"
+	monitoringpb "google.golang.org/genproto/googleapis/monitoring/v3"
+	"google.golang.org/grpc/codes"
+	"google.golang.org/grpc/status"
+
+	flaggerv1 "github.com/fluxcd/flagger/pkg/apis/flagger/v1beta1"
+)
+
+type StackDriverProvider struct {
+	client  *monitoring.QueryClient
+	project string
+}
+
+// NewStackDriverProvider takes a provider spec and credential map and
+// returns a StackDriverProvider ready to execute queries against the
+// Cloud Monitoring API
+func NewStackDriverProvider(provider flaggerv1.MetricTemplateProvider,
+	credentials map[string][]byte,
+) (*StackDriverProvider, error) {
+	stackd := &StackDriverProvider{}
+	var saKey []byte
+	if provider.SecretRef != nil {
+		if project, ok := credentials["project"]; ok {
+			stackd.project = fmt.Sprintf("projects/%s", string(project))
+		} else {
+			return nil, fmt.Errorf("%s credentials does not contain a project id", provider.Type)
+		}
+
+		if cred, ok := credentials["serviceAccountKey"]; ok {
+			saKey = cred
+		}
+	}
+
+	var client *monitoring.QueryClient
+	var err error
+	ctx := context.Background()
+
+	if saKey != nil {
+		client, err = monitoring.NewQueryClient(ctx, option.WithCredentialsJSON(saKey))
+	} else {
+		client, err = monitoring.NewQueryClient(ctx)
+	}
+
+	if err != nil {
+		return nil, err
+	}
+
+	stackd.client = client
+	return stackd, nil
+}
+
+// RunQuery executes Monitoring Query Language(MQL) queries against the
+// Cloud Monitoring API
+func (s *StackDriverProvider) RunQuery(query string) (float64, error) {
+	ctx := context.Background()
+	req := &monitoringpb.QueryTimeSeriesRequest{
+		Name:  s.project,
+		Query: query,
+	}
+
+	it := s.client.QueryTimeSeries(ctx, req)
+
+	resp, err := it.Next()
+	if err == iterator.Done {
+		return 0, fmt.Errorf("invalid response: %s: %w", resp, ErrNoValuesFound)
+	}
+
+	if err != nil {
+		if s, ok := status.FromError(err); ok {
+			errStr := s.Message()
+			for _, d := range s.Proto().Details {
+				errStr = errStr + " Error Detail: " + d.String()
+			}
+
+			return 0, fmt.Errorf("error requesting stackdriver: %s", err)
+		}
+		return 0, fmt.Errorf("error requesting stackdriver: %s", err)
+	}
+
+	pointData := resp.PointData
+	if len(pointData) < 1 {
+		return 0, fmt.Errorf("invalid response: %s: %w", resp.String(), ErrNoValuesFound)
+	}
+
+	values := resp.PointData[0].Values
+	if len(values) < 1 {
+		return 0, fmt.Errorf("invalid response: %s: %w", resp.String(), ErrNoValuesFound)
+	}
+
+	return values[0].GetDoubleValue(), nil
+}
+
+// IsOnline calls QueryTimeSeries method with the empty query
+// and returns an error if the returned status code is NOT grpc.InvalidArgument.
+// For example, if the flagger does not the authorization scope `https://www.googleapis.com/auth/monitoring.read`,
+// the returned status code would be grpc.PermissionDenied
+func (s *StackDriverProvider) IsOnline() (bool, error) {
+	ctx := context.Background()
+	req := &monitoringpb.QueryTimeSeriesRequest{
+		Name:  s.project,
+		Query: "",
+	}
+
+	it := s.client.QueryTimeSeries(ctx, req)
+
+	_, err := it.Next()
+	if err == nil {
+		return true, nil
+	}
+
+	stat, ok := status.FromError(err)
+	if !ok {
+		return false, fmt.Errorf("unexpected error: %s", err)
+	}
+
+	if stat.Code() != codes.InvalidArgument {
+		return false, fmt.Errorf("unexpected status code: %s", stat.Code().String())
+	}
+
+	return true, nil
+}