metrics: Report reason as a label on the conditions metric

Conditions are read by telemetry and the reason incentivizes teams
to have good reasons on their operators. Should not increase
cardinality significantly, but instead provide better insight.

Author: smarterclayton

docs/dev/metrics.md

@@ -31,10 +31,10 @@ Metrics about cluster operators:
 ```
 # HELP cluster_operator_conditions Report the conditions for active cluster operators. 0 is False and 1 is True.
 # TYPE cluster_operator_conditions gauge
-cluster_operator_conditions{condition="Available",name="version",namespace="openshift-cluster-version"} 1
-cluster_operator_conditions{condition="Degraded",name="version",namespace="openshift-cluster-version"} 0
-cluster_operator_conditions{condition="Progressing",name="version",namespace="openshift-cluster-version"} 0
-cluster_operator_conditions{condition="RetrievedUpdates",name="version",namespace="openshift-cluster-version"} 0
+cluster_operator_conditions{condition="Available",name="version",namespace="openshift-cluster-version",reason="Happy"} 1
+cluster_operator_conditions{condition="Degraded",name="version",namespace="openshift-cluster-version",reason=""} 0
+cluster_operator_conditions{condition="Progressing",name="version",namespace="openshift-cluster-version",reason=""} 0
+cluster_operator_conditions{condition="RetrievedUpdates",name="version",namespace="openshift-cluster-version",reason=""} 0
 # HELP cluster_operator_up Reports key highlights of the active cluster operators.
 # TYPE cluster_operator_up gauge
 cluster_operator_up{name="version",namespace="openshift-cluster-version",version="4.0.1"} 1

pkg/cvo/metrics.go

@@ -70,7 +70,7 @@ started.
 		clusterOperatorConditions: prometheus.NewGaugeVec(prometheus.GaugeOpts{
 			Name: "cluster_operator_conditions",
 			Help: "Report the conditions for active cluster operators. 0 is False and 1 is True.",
-		}, []string{"name", "condition"}),
+		}, []string{"name", "condition", "reason"}),
 		clusterOperatorConditionTransitions: prometheus.NewGaugeVec(prometheus.GaugeOpts{
 			Name: "cluster_operator_condition_transitions",
 			Help: "Reports the number of times that a condition on a cluster operator changes status",
@@ -122,7 +122,7 @@ func (m *operatorMetrics) Describe(ch chan<- *prometheus.Desc) {
 	ch <- m.version.WithLabelValues("", "", "").Desc()
 	ch <- m.availableUpdates.WithLabelValues("", "").Desc()
 	ch <- m.clusterOperatorUp.WithLabelValues("", "").Desc()
-	ch <- m.clusterOperatorConditions.WithLabelValues("", "").Desc()
+	ch <- m.clusterOperatorConditions.WithLabelValues("", "", "").Desc()
 	ch <- m.clusterOperatorConditionTransitions.WithLabelValues("", "").Desc()
 }
 
@@ -227,7 +227,7 @@ func (m *operatorMetrics) Collect(ch chan<- prometheus.Metric) {
 			if condition.Status == configv1.ConditionUnknown {
 				continue
 			}
-			g := m.clusterOperatorConditions.WithLabelValues(op.Name, string(condition.Type))
+			g := m.clusterOperatorConditions.WithLabelValues(op.Name, string(condition.Type), string(condition.Reason))
 			if condition.Status == configv1.ConditionTrue {
 				g.Set(1)
 			} else {

pkg/cvo/metrics_test.go

@@ -161,7 +161,7 @@ func Test_operatorMetrics_Collect(t *testing.T) {
 							Status: configv1.ClusterOperatorStatus{
 								Conditions: []configv1.ClusterOperatorStatusCondition{
 									{Type: configv1.OperatorAvailable, Status: configv1.ConditionTrue},
-									{Type: configv1.ClusterStatusConditionType("Custom"), Status: configv1.ConditionFalse},
+									{Type: configv1.ClusterStatusConditionType("Custom"), Status: configv1.ConditionFalse, Reason: "CustomReason"},
 									{Type: configv1.ClusterStatusConditionType("Unknown"), Status: configv1.ConditionUnknown},
 								},
 							},
@@ -175,8 +175,8 @@ func Test_operatorMetrics_Collect(t *testing.T) {
 				}
 				expectMetric(t, metrics[0], 0, map[string]string{"type": "current", "version": "", "image": ""})
 				expectMetric(t, metrics[1], 1, map[string]string{"name": "test", "version": ""})
-				expectMetric(t, metrics[2], 1, map[string]string{"name": "test", "condition": "Available"})
-				expectMetric(t, metrics[3], 0, map[string]string{"name": "test", "condition": "Custom"})
+				expectMetric(t, metrics[2], 1, map[string]string{"name": "test", "condition": "Available", "reason": ""})
+				expectMetric(t, metrics[3], 0, map[string]string{"name": "test", "condition": "Custom", "reason": "CustomReason"})
 			},
 		},
 		{