[Feature] Add SpecPropagated condition (#1212)

Author: ajanikow

CHANGELOG.md

@@ -2,6 +2,7 @@
 
 ## [master](https://github.com/arangodb/kube-arangodb/tree/master) (N/A)
 - (Bugfix) Remove PDBs if group count is 0
+- (Feature) Add SpecPropagated condition
 
 ## [1.2.22](https://github.com/arangodb/kube-arangodb/tree/1.2.22) (2022-12-13)
 - (Bugfix) Do not manage ports in managed ExternalAccess mode

docs/generated/metrics/README.md

@@ -24,6 +24,7 @@
 |                   [arangodb_operator_rebalancer_moves_succeeded](./arangodb_operator_rebalancer_moves_succeeded.md)                   | arangodb_operator |  rebalancer  | Counter | Define how many moves succeeded                                                       |
 |          [arangodb_operator_resources_arangodeployment_accepted](./arangodb_operator_resources_arangodeployment_accepted.md)          | arangodb_operator |  resources   |  Gauge  | Defines if ArangoDeployment has been accepted                                         |
 |  [arangodb_operator_resources_arangodeployment_immutable_errors](./arangodb_operator_resources_arangodeployment_immutable_errors.md)  | arangodb_operator |  resources   | Counter | Counter for deployment immutable errors                                               |
+|        [arangodb_operator_resources_arangodeployment_propagated](./arangodb_operator_resources_arangodeployment_propagated.md)        | arangodb_operator |  resources   |  Gauge  | Defines if ArangoDeployment Spec is propagated                                        |
 |   [arangodb_operator_resources_arangodeployment_status_restores](./arangodb_operator_resources_arangodeployment_status_restores.md)   | arangodb_operator |  resources   | Counter | Counter for deployment status restored                                                |
 |          [arangodb_operator_resources_arangodeployment_uptodate](./arangodb_operator_resources_arangodeployment_uptodate.md)          | arangodb_operator |  resources   |  Gauge  | Defines if ArangoDeployment is uptodate                                               |
 | [arangodb_operator_resources_arangodeployment_validation_errors](./arangodb_operator_resources_arangodeployment_validation_errors.md) | arangodb_operator |  resources   | Counter | Counter for deployment validation errors                                              |

docs/generated/metrics/arangodb_operator_resources_arangodeployment_propagated.md

@@ -0,0 +1,12 @@
+# arangodb_operator_resources_arangodeployment_propagated (Gauge)
+
+## Description
+
+Defines if ArangoDeployment Spec is propagated
+
+## Labels
+
+|   Label   | Description          |
+|:---------:|:---------------------|
+| namespace | Deployment Namespace |
+|   name    | Deployment Name      |

internal/metrics.yaml

@@ -186,6 +186,15 @@ namespaces:
             description: "Deployment Namespace"
           - key: name
             description: "Deployment Name"
+      arangodeployment_propagated:
+        shortDescription: "Defines if ArangoDeployment Spec is propagated"
+        description: "Defines if ArangoDeployment Spec is propagated"
+        type: "Gauge"
+        labels:
+          - key: namespace
+            description: "Deployment Namespace"
+          - key: name
+            description: "Deployment Name"
       arangodeployment_uptodate:
         shortDescription: "Defines if ArangoDeployment is uptodate"
         description: "Defines if ArangoDeployment is uptodate"

pkg/apis/deployment/v1/conditions.go

@@ -70,6 +70,8 @@ const (
 	ConditionTypeUpToDate ConditionType = "UpToDate"
 	// ConditionTypeSpecAccepted indicates that the deployment spec has been accepted.
 	ConditionTypeSpecAccepted ConditionType = "SpecAccepted"
+	// ConditionTypeSpecPropagated indicates that the deployment has been at least once UpToDate after spec acceptance.
+	ConditionTypeSpecPropagated ConditionType = "SpecPropagated"
 	// ConditionTypeMarkedToRemove indicates that the member is marked to be removed.
 	ConditionTypeMarkedToRemove ConditionType = "MarkedToRemove"
 	// ConditionTypeScaleDownCandidate indicates that the member will be picked in ScaleDown operaion.

pkg/apis/deployment/v2alpha1/conditions.go

@@ -70,6 +70,8 @@ const (
 	ConditionTypeUpToDate ConditionType = "UpToDate"
 	// ConditionTypeSpecAccepted indicates that the deployment spec has been accepted.
 	ConditionTypeSpecAccepted ConditionType = "SpecAccepted"
+	// ConditionTypeSpecPropagated indicates that the deployment has been at least once UpToDate after spec acceptance.
+	ConditionTypeSpecPropagated ConditionType = "SpecPropagated"
 	// ConditionTypeMarkedToRemove indicates that the member is marked to be removed.
 	ConditionTypeMarkedToRemove ConditionType = "MarkedToRemove"
 	// ConditionTypeScaleDownCandidate indicates that the member will be picked in ScaleDown operaion.

pkg/deployment/deployment_inspector.go

@@ -143,6 +143,7 @@ func (d *Deployment) inspectDeployment(lastInterval util.Interval) util.Interval
 		d.currentObject = updated
 
 		d.metrics.Deployment.Accepted = updated.Status.Conditions.IsTrue(api.ConditionTypeSpecAccepted)
+		d.metrics.Deployment.Propagated = updated.Status.Conditions.IsTrue(api.ConditionTypeSpecPropagated)
 		d.metrics.Deployment.UpToDate = updated.Status.Conditions.IsTrue(api.ConditionTypeUpToDate)
 
 		// Is the deployment in failed state, if so, give up.
@@ -231,6 +232,16 @@ func (d *Deployment) inspectDeploymentWithError(ctx context.Context, lastInterva
 	if !status.Conditions.IsTrue(api.ConditionTypeSpecAccepted) {
 		condition, exists := status.Conditions.Get(api.ConditionTypeUpToDate)
 		if !exists || condition.IsTrue() {
+			propagatedCondition, propagatedExists := status.Conditions.Get(api.ConditionTypeSpecPropagated)
+			if !propagatedExists || propagatedCondition.IsTrue() {
+				if err = d.updateConditionWithHash(ctx, api.ConditionTypeSpecPropagated, false, "Spec Changed", "Spec Object changed. Waiting until spec will be applied", ""); err != nil {
+					return minInspectionInterval, errors.Wrapf(err, "Unable to update SpecPropagated condition")
+
+				}
+
+				return minInspectionInterval, nil // Retry ASAP
+			}
+
 			if err = d.updateConditionWithHash(ctx, api.ConditionTypeUpToDate, false, "Spec Changed", "Spec Object changed. Waiting until plan will be applied", currentChecksum); err != nil {
 				return minInspectionInterval, errors.Wrapf(err, "Unable to update UpToDate condition")
 
@@ -374,6 +385,12 @@ func (d *Deployment) inspectDeploymentWithError(ctx context.Context, lastInterva
 		}
 	}
 
+	if status.Conditions.IsTrue(api.ConditionTypeUpToDate) && !status.Conditions.IsTrue(api.ConditionTypeSpecPropagated) {
+		if err = d.updateConditionWithHash(ctx, api.ConditionTypeSpecPropagated, true, "Spec is Propagated", "Spec is Propagated", ""); err != nil {
+			return minInspectionInterval, errors.Wrapf(err, "Unable to update SpecPropagated condition")
+		}
+	}
+
 	// Execute current step of scale/update plan
 	retrySoon, err := d.reconciler.ExecutePlan(ctx)
 	if err != nil {

pkg/deployment/metrics.go

@@ -39,7 +39,7 @@ type Metrics struct {
 	ArangodbOperatorEngineOpsAlerts int
 
 	Deployment struct {
-		Accepted, UpToDate bool
+		Accepted, UpToDate, Propagated bool
 	}
 }
 
@@ -62,6 +62,11 @@ func (d *Deployment) CollectMetrics(m metrics.PushMetric) {
 	} else {
 		m.Push(metric_descriptions.ArangodbOperatorResourcesArangodeploymentUptodateGauge(0, d.namespace, d.name))
 	}
+	if d.metrics.Deployment.Propagated {
+		m.Push(metric_descriptions.ArangodbOperatorResourcesArangodeploymentPropagatedGauge(1, d.namespace, d.name))
+	} else {
+		m.Push(metric_descriptions.ArangodbOperatorResourcesArangodeploymentPropagatedGauge(0, d.namespace, d.name))
+	}
 
 	if c := d.agencyCache; c != nil {
 		m.Push(metric_descriptions.ArangodbOperatorAgencyCachePresentGauge(1, d.namespace, d.name))