How to fix unable to compute new replica count for HPA: missing request metric in Kubernetes

What this error means

HPA performs a calculation: desiredReplicas = currentMetricValue / targetMetricValue * currentReplicas. When metric values are missing, NaN, or invalid, this calculation fails and HPA cannot determine the desired replica count. HPA will report status as "Unknown" and not change replica count.

Typical symptoms

• HPA status shows "Unknown" for desired replicas
• HPA events: "unable to compute new replica count"
• Metrics show as "unknown" in kubectl describe hpa
• Pod count does not change despite metric changes
• Missing metric values in HPA calculation
• Metric returns 0 or NaN for specific pods
• Inconsistent metric data from some replicas

Common causes

1. Pod missing from some metric queries (new pods)
2. Metric returns partial results (some pods missing)
3. Pod metrics include pods not in Service selector
4. Metric aggregation error (avg, sum, etc.)
5. Resource request missing from pod spec
6. Metric value is NaN or 0 for all pods
7. Custom metric not properly registered

How to fix "unable to compute new replica count for HPA: missing request metric"

Advanced notes

Frequently asked questions

What does "unable to compute new replica count for HPA: missing request metric" mean?

HPA performs a calculation: desiredReplicas = currentMetricValue / targetMetricValue * currentReplicas. When metric values are missing, NaN, or invalid, this calculation fails and HPA cannot determine the desired replica count. HPA will report status as "Unknown" and not change replica count.

How do I fix "unable to compute new replica count for HPA: missing request metric"?

HorizontalPodAutoscaler cannot calculate desired replica count due to missing metrics, invalid metric values, or computation errors. Scaling decisions fail.

What causes "unable to compute new replica count for HPA: missing request metric"?

Pod missing from some metric queries (new pods) Metric returns partial results (some pods missing) Pod metrics include pods not in Service selector

Is the "unable to compute new replica count for HPA: missing request metric" error serious?

Yes. This is a high-severity error that usually blocks work until it is fixed, but the steps above resolve it.

Related errors

• unable to get metrics for resource cpu: no metrics returned from resource metrics API

  How to fix "HorizontalPodAutoscaler unable to scale" in Kubernetes

  MEDIUM

• unable to compute replica count

  How to fix "unable to compute replica count" in Kubernetes HPA

  MEDIUM

• Failed to connect to server: connection refused (HTTP/2)

  How to fix "HTTP/2 connection refused" error in Kubernetes

  HIGH

• ScaleTargetRef not found

  How to fix "ScaleTargetRef not found" in Kubernetes HPA

  HIGH

• upstream connect error or disconnect/reset before headers. reset reason: overflow

  How to fix "upstream connect error or disconnect/reset before headers" in Kubernetes

  HIGH

• Waiting for rollout to finish

  How to fix "Waiting for rollout to finish" in Kubernetes

  HIGH

Official resources & further reading

• 📚Kubernetes Official Documentation
• 📖
  Kubernetes Docs: Horizontal Pod Autoscaling(Official Docs)
• 📖
  Kubernetes Docs: Debugging Applications(Official Docs)
• 💬
  Stack Overflow: Kubernetes metrics server issues(Stack Overflow)