Impact If Worker Node Goes Down Where 'calico-kube-controllers', 'metrics-server', 'dashboard-metrics-scrape' and 'kubernetes-dashboard' Pods Are Running

Problem

What will be impact if worker node goes down where the below pods are running:

• calico-kube-controllers
• metrics-server-v0.5.0
• dashboard-metrics-scraper
• kubernetes-dashboard

Environment

• Platform9 Managed Kubernetes - v5.1 and Higher

Answer

calico-kube-controllers:

• Outside of calico being unable to sync any changes to the pod or namespace objects to apply policies, etc. would be impacted the most when the calico kube controller goes down.
• Garbage collects IP addresses. Automatically provisions host endpoints for Kubernetes nodes (from: https://projectcalico.docs.tigera.io/reference/kube-controllers/configuration)

kubernetes-dashboard and dashboard-metrics-scraper:

• The kubernetes dashboard will be unavailable in the first case
• The dashboard will not have the basic metrics that it usually displays.

metrics-server-v0.5.0:

• horizontal pod autoscalers
• vertical pod autoscalers
• kubectl top commands won't work.

Additional Information

These pods can be configured high availability mode by increasing the replicas.