Skip to content

🌱 Improve Machine remediation logs #11692

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
k8s-ci-robot merged 2 commits into from
Jan 20, 2025
Merged

🌱 Improve Machine remediation logs #11692

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
2a3b552
Improve Machine remediation logs
fabriziopandini Jan 16, 2025
5e3be89
Address comments
fabriziopandini Jan 17, 2025
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
12 changes: 7 additions & 5 deletions internal/controllers/machinehealthcheck/machinehealthcheck_controller.go
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -428,10 +428,11 @@ func (r *Reconciler) patchUnhealthyTargets(ctx context.Context, logger logr.Logg
// mark for remediation
errList := []error{}
for _, t := range unhealthy {
logger := logger.WithValues("Machine", klog.KObj(t.Machine), "Node", klog.KObj(t.Node))
condition := conditions.Get(t.Machine, clusterv1.MachineHealthCheckSucceededCondition)

if annotations.IsPaused(cluster, t.Machine) {
logger.Info("Machine has failed health check, but machine is paused so skipping remediation", "target", t.string(), "reason", condition.Reason, "message", condition.Message)
logger.Info("Machine has failed health check, but machine is paused so skipping remediation", "reason", condition.Reason, "message", condition.Message)
} else {
if m.Spec.RemediationTemplate != nil {
// If external remediation request already exists,
Expand Down Expand Up @@ -482,7 +483,7 @@ func (r *Reconciler) patchUnhealthyTargets(ctx context.Context, logger logr.Logg
// the same Machine, users are in charge of setting health checks and remediation properly.
to.SetName(t.Machine.Name)

logger.Info("Target has failed health check, creating an external remediation request", "remediation request name", to.GetName(), "target", t.string(), "reason", condition.Reason, "message", condition.Message)
logger.Info("Machine has failed health check, creating an external remediation request", "remediation request name", to.GetName(), "reason", condition.Reason, "message", condition.Message)
// Create the external clone.
if err := r.Client.Create(ctx, to); err != nil {
conditions.MarkFalse(m, clusterv1.ExternalRemediationRequestAvailableCondition, clusterv1.ExternalRemediationRequestCreationFailedReason, clusterv1.ConditionSeverityError, err.Error())
Expand All @@ -503,7 +504,7 @@ func (r *Reconciler) patchUnhealthyTargets(ctx context.Context, logger logr.Logg
Reason: clusterv1.MachineExternallyRemediatedWaitingForRemediationV1Beta2Reason,
})
} else if t.Machine.DeletionTimestamp.IsZero() { // Only setting the OwnerRemediated conditions when machine is not already in deletion.
logger.Info("Target has failed health check, marking for remediation", "target", t.string(), "reason", condition.Reason, "message", condition.Message)
logger.Info("Machine has failed health check, marking for remediation", "reason", condition.Reason, "message", condition.Message)
// NOTE: MHC is responsible for creating MachineOwnerRemediatedCondition if missing or to trigger another remediation if the previous one is completed;
// instead, if a remediation is in already progress, the remediation owner is responsible for completing the process and MHC should not overwrite the condition.
if !conditions.Has(t.Machine, clusterv1.MachineOwnerRemediatedCondition) || conditions.IsTrue(t.Machine, clusterv1.MachineOwnerRemediatedCondition) {
Expand Down Expand Up @@ -540,8 +541,9 @@ func (r *Reconciler) patchUnhealthyTargets(ctx context.Context, logger logr.Logg
t.Machine,
corev1.EventTypeNormal,
EventMachineMarkedUnhealthy,
"Machine %v has been marked as unhealthy",
t.string(),
"Machine %s has been marked as unhealthy by %s",
klog.KObj(t.Machine),
klog.KObj(t.MHC),
)
}
return errList
Expand Down
11 changes: 1 addition & 10 deletions internal/controllers/machinehealthcheck/machinehealthcheck_targets.go
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -64,15 +64,6 @@ type healthCheckTarget struct {
nodeMissing bool
}

func (t *healthCheckTarget) string() string {
return fmt.Sprintf("%s/%s/%s/%s",
t.MHC.GetNamespace(),
t.MHC.GetName(),
t.Machine.GetName(),
t.nodeName(),
)
}

// Get the node name if the target has a node.
func (t *healthCheckTarget) nodeName() string {
if t.Node != nil {
Expand Down Expand Up @@ -331,7 +322,7 @@ func (r *Reconciler) healthCheckTargets(targets []healthCheckTarget, logger logr
var healthy []healthCheckTarget

for _, t := range targets {
logger := logger.WithValues("target", t.string())
logger := logger.WithValues("Machine", klog.KObj(t.Machine), "Node", klog.KObj(t.Node))
logger.V(3).Info("Health checking target")
needsRemediation, nextCheck := t.needsRemediation(logger, timeoutForMachineToHaveNode)

Expand Down
Loading