Merge pull request #1270 from fpetkovski/indicate-no-storage

Bug 1978662: Set a degraded message when persistent storage is not configured

Author: openshift-merge-robot

Diff for: CHANGELOG.md

@@ -4,6 +4,7 @@
 
 - [#1241](https://github.com/openshift/cluster-monitoring-operator/pull/1241) Add config option to disable Grafana deployment.
 - [#1278](https://github.com/openshift/cluster-monitoring-operator/pull/1278) Add EnforcedTargetLimit option for user-workload Prometheus.
+- [#1270](https://github.com/openshift/cluster-monitoring-operator/pull/1270) Show a message in the degraded condition when Platform Monitoring Prometheus runs without persistent storage. 
 
 ## 4.8
 

Diff for: pkg/client/status_reporter.go

@@ -29,6 +29,8 @@ import (
 const (
 	unavailableMessage string = "Rollout of the monitoring stack failed and is degraded. Please investigate the degraded status error."
 	asExpectedReason   string = "AsExpected"
+	StorageNotConfiguredMessage = "Prometheus is running without persistent storage which can lead to data loss during upgrades and cluster disruptions. Please refer to the official documentation to see how to configure storage for Prometheus: https://docs.openshift.com/container-platform/4.8/monitoring/configuring-the-monitoring-stack.html"
+	StorageNotConfiguredReason  = "PrometheusDataPersistenceNotConfigured"
 )
 
 type StatusReporter struct {
@@ -67,7 +69,7 @@ func (r *StatusReporter) relatedObjects() []v1.ObjectReference {
 	}
 }
 
-func (r *StatusReporter) SetDone() error {
+func (r *StatusReporter) SetDone(degradedConditionMessage string, degradedConditionReason string) error {
 	co, err := r.client.Get(r.ctx, r.clusterOperatorName, metav1.GetOptions{})
 	if apierrors.IsNotFound(err) {
 		co = r.newClusterOperator()
@@ -82,7 +84,7 @@ func (r *StatusReporter) SetDone() error {
 	conditions := newConditions(co.Status, r.version, time)
 	conditions.setCondition(v1.OperatorAvailable, v1.ConditionTrue, "Successfully rolled out the stack.", "RollOutDone", time)
 	conditions.setCondition(v1.OperatorProgressing, v1.ConditionFalse, "", "", time)
-	conditions.setCondition(v1.OperatorDegraded, v1.ConditionFalse, "", "", time)
+	conditions.setCondition(v1.OperatorDegraded, v1.ConditionFalse, degradedConditionMessage, degradedConditionReason, time)
 	conditions.setCondition(v1.OperatorUpgradeable, v1.ConditionTrue, "", asExpectedReason, time)
 	co.Status.Conditions = conditions.entries()
 

Diff for: pkg/client/status_reporter_test.go

@@ -113,7 +113,7 @@ func TestStatusReporterSetDone(t *testing.T) {
 				w(mock)
 			}
 
-			got := sr.SetDone()
+			got := sr.SetDone("", "")
 
 			for _, check := range tc.check {
 				if err := check(mock, got); err != nil {

Diff for: pkg/manifests/config.go

@@ -38,6 +38,19 @@ type Config struct {
 	UserWorkloadConfiguration      *UserWorkloadConfiguration      `json:"-"`
 }
 
+func (c Config) IsStorageConfigured() bool {
+	if c.ClusterMonitoringConfiguration == nil {
+		return false
+	}
+
+	prometheusK8sConfig := c.ClusterMonitoringConfiguration.PrometheusK8sConfig
+	if prometheusK8sConfig == nil {
+		return false
+	}
+
+	return prometheusK8sConfig.VolumeClaimTemplate != nil
+}
+
 type ClusterMonitoringConfiguration struct {
 	PrometheusOperatorConfig *PrometheusOperatorConfig    `json:"prometheusOperator"`
 	PrometheusK8sConfig      *PrometheusK8sConfig         `json:"prometheusK8s"`

Diff for: pkg/operator/operator.go

@@ -481,9 +481,14 @@ func (o *Operator) sync(key string) error {
 		return err
 	}
 
+	var degradedConditionMessage, degradedConditionReason string
+	if !config.IsStorageConfigured() {
+		degradedConditionMessage = client.StorageNotConfiguredMessage
+		degradedConditionReason = client.StorageNotConfiguredReason
+	}
 	klog.Info("Updating ClusterOperator status to done.")
 	o.failedReconcileAttempts = 0
-	err = o.client.StatusReporter().SetDone()
+	err = o.client.StatusReporter().SetDone(degradedConditionMessage, degradedConditionReason)
 	if err != nil {
 		klog.Errorf("error occurred while setting status to done: %v", err)
 	}