What causes "desiring %d replicas but %d ready" in derailed/k9s?

Wait if a rollout is in flight: kubectl rollout status ds/

desiring %d replicas but %d ready

Error message

desiring %d replicas but %d ready

What it means

DaemonSet.diagnose compares status.desiredNumberScheduled with status.numberReady; a mismatch yields 'desiring N replicas but M ready'. It is the health indicator for DaemonSet rows: nodes where the daemon pod is not (yet) ready.

Source

Thrown at internal/render/ds.go:89

		ds.Name,
		computeVulScore(ds.Namespace, ds.Labels, &ds.Spec.Template.Spec),
		strconv.Itoa(int(ds.Status.DesiredNumberScheduled)),
		strconv.Itoa(int(ds.Status.CurrentNumberScheduled)),
		strconv.Itoa(int(ds.Status.NumberReady)),
		strconv.Itoa(int(ds.Status.UpdatedNumberScheduled)),
		strconv.Itoa(int(ds.Status.NumberAvailable)),
		mapToStr(ds.Labels),
		AsStatus(d.diagnose(ds.Status.DesiredNumberScheduled, ds.Status.NumberReady)),
		ToAge(ds.GetCreationTimestamp()),
	}

	return nil
}

// Happy returns true if resource is happy, false otherwise.
func (DaemonSet) diagnose(d, r int32) error {
	if d != r {
		return fmt.Errorf("desiring %d replicas but %d ready", d, r)
	}
	return nil
}

View on GitHub (pinned to 2d3ccc6ba2)

Solutions

Wait if a rollout is in flight: kubectl rollout status ds/<name>
Find not-ready pods: kubectl get pods -l <selector> -o wide | grep -v Running/Ready, then describe the failing ones
Add missing tolerations/affinity for tainted nodes, or fix resources
Check kubelet/containerd on nodes whose daemon pods stay pending

Example fix

# before: daemon pods not scheduled on control-plane
# tolerations missing
# after
tolerations:
- key: node-role.kubernetes.io/control-plane
  operator: Exists
  effect: NoSchedule

Defensive patterns

Strategy: fallback

Validate before calling

if ds.Status.DesiredNumberScheduled == ds.Status.NumberReady {
	status = "ok" // skip diagnose
}

Try / catch

Use the error as row status; when desired-ready persists across refreshes, escalate to per-node pod inspection.

Prevention

Add tolerations for all node roles your daemon must run on
Pre-pull daemon images or use registry mirrors on large fleets
Alert on per-node daemon gaps rather than aggregate mismatch alone

When it happens

Trigger: Daemon rollout in progress on a large fleet; pods unschedulable on some nodes (taints, resources); CNI/node-exporter daemons crashing on specific kernels; newly joined nodes still pulling images.

Common situations: Upgrading kube-proxy/node-exporter DaemonSets; nodes with incompatible taints missing tolerations; slow registries delaying image pulls; disk pressure evicting daemon pods.

Related errors

AI-assisted analysis of derailed/k9s@2d3ccc6ba2 (2026-08-15). Data as JSON: /api/errors/1788ff839994118d. Report an issue: GitHub.