Skip to content
Open
Show file tree
Hide file tree
Changes from all commits
Commits
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
2 changes: 2 additions & 0 deletions internal/controller/node_controller.go
Original file line number Diff line number Diff line change
Expand Up @@ -148,6 +148,8 @@ func (r *RuleReadinessController) processNodeAgainstAllRules(ctx context.Context
"node", node.Name, "rule", rule.Name)
// Continue with other rules even if one fails
r.recordNodeFailure(rule, node.Name, "EvaluationError", err.Error())
} else {
r.clearNodeFailure(rule, node.Name)
}

// Persist the rule status
Expand Down
15 changes: 14 additions & 1 deletion internal/controller/nodereadinessrule_controller.go
Original file line number Diff line number Diff line change
Expand Up @@ -277,6 +277,8 @@ func (r *RuleReadinessController) processAllNodesForRule(ctx context.Context, ru
log.Error(err, "Failed to evaluate node for rule", "rule", rule.Name, "node", node.Name)
r.recordNodeFailure(rule, node.Name, "EvaluationError", err.Error())
metrics.Failures.WithLabelValues(rule.Name, "EvaluationError").Inc()
} else {
r.clearNodeFailure(rule, node.Name)
}
}
}
Expand Down Expand Up @@ -422,7 +424,7 @@ func (r *RuleReadinessController) getApplicableRulesForNode(ctx context.Context,

for _, rule := range r.ruleCache {
if r.ruleAppliesTo(ctx, rule, node) {
applicableRules = append(applicableRules, rule)
applicableRules = append(applicableRules, rule.DeepCopy())
}
}

Expand Down Expand Up @@ -706,3 +708,14 @@ func (r *RuleReadinessController) getPreviousNodeEvaluation(rule *readinessv1alp
}
return nil
}

// clearNodeFailure removes any failure record for a specific node from the rule status.
func (r *RuleReadinessController) clearNodeFailure(rule *readinessv1alpha1.NodeReadinessRule, nodeName string) {
var failedNodes []readinessv1alpha1.NodeFailure
for _, failure := range rule.Status.FailedNodes {
if failure.NodeName != nodeName {
failedNodes = append(failedNodes, failure)
}
}
rule.Status.FailedNodes = failedNodes
}