Improve error messaging around bootstrap extension provisioning

Author: Cecile Robert-Michon

azure/errors.go

@@ -89,11 +89,9 @@ func (t ReconcileError) Error() string {
 	}
 	switch t.errorType {
 	case TransientErrorType:
-		return fmt.Sprintf("reconcile error occurred that can be recovered. Object will be requeued after %s "+
-			"The actual error is: %s", t.requestAfter.String(), errStr)
+		return fmt.Sprintf("transient reconcile error occurred: %s. Object will be requeued after %s", errStr, t.requestAfter.String())
 	case TerminalErrorType:
-		return fmt.Sprintf("reconcile error occurred that cannot be recovered. Object will not be requeued. "+
-			"The actual error is: %s", errStr)
+		return fmt.Sprintf("reconcile error that cannot be recovered occurred: %s. Object will not be requeued", errStr)
 	default:
 		return fmt.Sprintf("reconcile error occurred with unknown recovery type. The actual error is: %s", errStr)
 	}

azure/scope/machine.go

@@ -379,11 +379,11 @@ func (m *MachineScope) SetBootstrapConditions(provisioningState string, extensio
 	case infrav1.Creating:
 		m.V(4).Info("extension provisioning state is creating", "vm extension", extensionName, "virtual machine", m.Name())
 		conditions.MarkFalse(m.AzureMachine, infrav1.BootstrapSucceededCondition, infrav1.BootstrapInProgressReason, clusterv1.ConditionSeverityInfo, "")
-		return azure.WithTransientError(errors.New("extension still provisioning"), 30*time.Second)
+		return azure.WithTransientError(errors.New("extension is still in provisioning state. This likely means that bootstrapping has not yet completed on the VM"), 30*time.Second)
 	case infrav1.Failed:
 		m.V(4).Info("extension provisioning state is failed", "vm extension", extensionName, "virtual machine", m.Name())
 		conditions.MarkFalse(m.AzureMachine, infrav1.BootstrapSucceededCondition, infrav1.BootstrapFailedReason, clusterv1.ConditionSeverityError, "")
-		return azure.WithTerminalError(errors.New("extension state failed"))
+		return azure.WithTerminalError(errors.New("extension state failed. This likely means the Kubernetes node bootstrapping process failed or timed out. Check VM boot diagnostics logs to learn more"))
 	default:
 		return nil
 	}

azure/scope/machinepool.go

@@ -447,11 +447,11 @@ func (m *MachinePoolScope) SetBootstrapConditions(provisioningState string, exte
 	case infrav1.Creating:
 		m.V(4).Info("extension provisioning state is creating", "vm extension", extensionName, "scale set", m.Name())
 		conditions.MarkFalse(m.AzureMachinePool, infrav1.BootstrapSucceededCondition, infrav1.BootstrapInProgressReason, clusterv1.ConditionSeverityInfo, "")
-		return azure.WithTransientError(errors.New("extension still provisioning"), 30*time.Second)
+		return azure.WithTransientError(errors.New("extension is still in provisioning state. This likely means that bootstrapping has not yet completed on the VM"), 30*time.Second)
 	case infrav1.Failed:
 		m.V(4).Info("extension provisioning state is failed", "vm extension", extensionName, "scale set", m.Name())
 		conditions.MarkFalse(m.AzureMachinePool, infrav1.BootstrapSucceededCondition, infrav1.BootstrapFailedReason, clusterv1.ConditionSeverityError, "")
-		return azure.WithTerminalError(errors.New("extension state failed"))
+		return azure.WithTerminalError(errors.New("extension state failed. This likely means the Kubernetes node bootstrapping process failed or timed out. Check VM boot diagnostics logs to learn more"))
 	default:
 		return nil
 	}

azure/scope/machinepool_test.go

@@ -116,7 +116,7 @@ func TestMachinePoolScope_SetBootstrapConditions(t *testing.T) {
 				return string(infrav1.Creating), "bazz"
 			},
 			Verify: func(g *WithT, amp *infrav1exp.AzureMachinePool, err error) {
-				g.Expect(err).To(MatchError("reconcile error occurred that can be recovered. Object will be requeued after 30s The actual error is: extension still provisioning"))
+				g.Expect(err).To(MatchError("transient reconcile error occurred: extension is still in provisioning state. This likely means that bootstrapping has not yet completed on the VM. Object will be requeued after 30s"))
 				g.Expect(conditions.IsFalse(amp, infrav1.BootstrapSucceededCondition))
 				g.Expect(conditions.GetReason(amp, infrav1.BootstrapSucceededCondition)).To(Equal(infrav1.BootstrapInProgressReason))
 				severity := conditions.GetSeverity(amp, infrav1.BootstrapSucceededCondition)
@@ -130,7 +130,7 @@ func TestMachinePoolScope_SetBootstrapConditions(t *testing.T) {
 				return string(infrav1.Failed), "buzz"
 			},
 			Verify: func(g *WithT, amp *infrav1exp.AzureMachinePool, err error) {
-				g.Expect(err).To(MatchError("reconcile error occurred that cannot be recovered. Object will not be requeued. The actual error is: extension state failed"))
+				g.Expect(err).To(MatchError("reconcile error that cannot be recovered occurred: extension state failed. This likely means the Kubernetes node bootstrapping process failed or timed out. Check VM boot diagnostics logs to learn more. Object will not be requeued"))
 				g.Expect(conditions.IsFalse(amp, infrav1.BootstrapSucceededCondition))
 				g.Expect(conditions.GetReason(amp, infrav1.BootstrapSucceededCondition)).To(Equal(infrav1.BootstrapFailedReason))
 				severity := conditions.GetSeverity(amp, infrav1.BootstrapSucceededCondition)

azure/services/scalesets/scalesets_test.go

@@ -402,7 +402,7 @@ func TestReconcileVMSS(t *testing.T) {
 		},
 		{
 			name:          "creating a vmss with encryption at host enabled for unsupported VM type fails",
-			expectedError: "reconcile error occurred that cannot be recovered. Object will not be requeued. The actual error is: encryption at host is not supported for VM type VM_SIZE",
+			expectedError: "reconcile error that cannot be recovered occurred: encryption at host is not supported for VM type VM_SIZE. Object will not be requeued",
 			expect: func(g *WithT, s *mock_scalesets.MockScaleSetScopeMockRecorder, m *mock_scalesets.MockClientMockRecorder) {
 				s.ScaleSetSpec().Return(azure.ScaleSetSpec{
 					Name:            defaultVMSSName,
@@ -444,7 +444,7 @@ func TestReconcileVMSS(t *testing.T) {
 		},
 		{
 			name:          "less than 2 vCPUs",
-			expectedError: "reconcile error occurred that cannot be recovered. Object will not be requeued. The actual error is: vm size should be bigger or equal to at least 2 vCPUs",
+			expectedError: "reconcile error that cannot be recovered occurred: vm size should be bigger or equal to at least 2 vCPUs. Object will not be requeued",
 			expect: func(g *WithT, s *mock_scalesets.MockScaleSetScopeMockRecorder, m *mock_scalesets.MockClientMockRecorder) {
 				s.ScaleSetSpec().Return(azure.ScaleSetSpec{
 					Name:       defaultVMSSName,
@@ -456,7 +456,7 @@ func TestReconcileVMSS(t *testing.T) {
 		},
 		{
 			name:          "Memory is less than 2Gi",
-			expectedError: "reconcile error occurred that cannot be recovered. Object will not be requeued. The actual error is: vm memory should be bigger or equal to at least 2Gi",
+			expectedError: "reconcile error that cannot be recovered occurred: vm memory should be bigger or equal to at least 2Gi. Object will not be requeued",
 			expect: func(g *WithT, s *mock_scalesets.MockScaleSetScopeMockRecorder, m *mock_scalesets.MockClientMockRecorder) {
 				s.ScaleSetSpec().Return(azure.ScaleSetSpec{
 					Name:       defaultVMSSName,
@@ -468,7 +468,7 @@ func TestReconcileVMSS(t *testing.T) {
 		},
 		{
 			name:          "failed to get SKU",
-			expectedError: "reconcile error occurred that cannot be recovered. Object will not be requeued. The actual error is: failed to get SKU INVALID_VM_SIZE in compute api: resource sku with name 'INVALID_VM_SIZE' and category 'virtualMachines' not found in location 'test-location'",
+			expectedError: "reconcile error that cannot be recovered occurred: failed to get SKU INVALID_VM_SIZE in compute api: resource sku with name 'INVALID_VM_SIZE' and category 'virtualMachines' not found in location 'test-location'. Object will not be requeued",
 			expect: func(g *WithT, s *mock_scalesets.MockScaleSetScopeMockRecorder, m *mock_scalesets.MockClientMockRecorder) {
 				s.ScaleSetSpec().Return(azure.ScaleSetSpec{
 					Name:       defaultVMSSName,

azure/services/virtualmachines/virtualmachines_test.go

@@ -1117,7 +1117,7 @@ func TestReconcileVM(t *testing.T) {
 				m.Get(gomockinternal.AContext(), "my-rg", "my-vm").
 					Return(compute.VirtualMachine{}, autorest.NewErrorWithResponse("", "", &http.Response{StatusCode: 404}, "Not found"))
 			},
-			ExpectedError: "reconcile error occurred that cannot be recovered. Object will not be requeued. The actual error is: encryption at host is not supported for VM type Standard_D2v3",
+			ExpectedError: "reconcile error that cannot be recovered occurred: encryption at host is not supported for VM type Standard_D2v3. Object will not be requeued",
 			SetupSKUs: func(svc *Service) {
 				skus := []compute.ResourceSku{
 					{
@@ -1252,7 +1252,7 @@ func TestReconcileVM(t *testing.T) {
 				m.Get(gomockinternal.AContext(), "my-rg", "my-vm").
 					Return(compute.VirtualMachine{}, autorest.NewErrorWithResponse("", "", &http.Response{StatusCode: 404}, "Not found"))
 			},
-			ExpectedError: "reconcile error occurred that cannot be recovered. Object will not be requeued. The actual error is: vm size should be bigger or equal to at least 2 vCPUs",
+			ExpectedError: "reconcile error that cannot be recovered occurred: vm size should be bigger or equal to at least 2 vCPUs. Object will not be requeued",
 			SetupSKUs: func(svc *Service) {
 				skus := []compute.ResourceSku{
 					{
@@ -1319,7 +1319,7 @@ func TestReconcileVM(t *testing.T) {
 				m.Get(gomockinternal.AContext(), "my-rg", "my-vm").
 					Return(compute.VirtualMachine{}, autorest.NewErrorWithResponse("", "", &http.Response{StatusCode: 404}, "Not found"))
 			},
-			ExpectedError: "reconcile error occurred that cannot be recovered. Object will not be requeued. The actual error is: vm memory should be bigger or equal to at least 2Gi",
+			ExpectedError: "reconcile error that cannot be recovered occurred: vm memory should be bigger or equal to at least 2Gi. Object will not be requeued",
 			SetupSKUs: func(svc *Service) {
 				skus := []compute.ResourceSku{
 					{
@@ -1389,7 +1389,7 @@ func TestReconcileVM(t *testing.T) {
 				m.Get(gomockinternal.AContext(), "my-rg", "my-vm").
 					Return(compute.VirtualMachine{}, autorest.NewErrorWithResponse("", "", &http.Response{StatusCode: 404}, "Not found"))
 			},
-			ExpectedError: "reconcile error occurred that cannot be recovered. Object will not be requeued. The actual error is: vm size Standard_D2v3 does not support ephemeral os. select a different vm size or disable ephemeral os",
+			ExpectedError: "reconcile error that cannot be recovered occurred: vm size Standard_D2v3 does not support ephemeral os. select a different vm size or disable ephemeral os. Object will not be requeued",
 			SetupSKUs: func(svc *Service) {
 				skus := []compute.ResourceSku{
 					{