Metrics Server and Kube-State-Metrics in Kubernetes.

 Metrics Server in Kubernetes

Metrics Server is a lightweight resource usage monitoring component in Kubernetes. It provides real-time CPU and memory metrics for nodes and pods, which are used by:

Does not store long-term data, only real-time values.

Only CPU & Memory is monitoring .

Horizontal Pod Autoscaler (HPA) – Auto-scales pods based on CPU/memory usage
kubectl top – View resource usage of pods and nodes
Custom Monitoring – Fetch live metrics via API

--kubelet-insecure-tls: Disable TLS verification when communicating with the kubelet (useful for self-signed certificates).

--kubelet-preferred-address-types: Specify the order of address types to use when connecting to the kubelet (e.g., InternalIP, Hostname).

 

# wget  https://github.com/kubernetes-sigs/metrics-server/releases/latest/download/components.yaml

# vim components.yaml

- --kubelet-insecure-tls

# kubectl apply -f components.yaml

 

# kubectl get pods -n kube-system | grep metrics-server

# kubectl logs -n kube-system -l k8s-app=metrics-server

# kubectl get pods -n kube-system | grep "metrics”

After installation, you can list the resources created by the Metrics Server:

kubectl get all -n kube-system | grep metrics-server

 

APIService:

An APIService named v1beta1.metrics.k8s.io registers the Metrics Server with the Kubernetes API.

# kubectl get apiservices

# kubectl top nodes

# kubectl top pods -A

 

These commands are used to check the health status of Kubernetes components, specifically the API server.

1.Readiness Probe Check (/readyz)

 Checks if the Kubernetes API server is ready to handle requests.

If the API server is not ready, it won’t accept new connections.

# kubectl get --raw /readyz

2.Liveness Probe Check (/livez)

Checks if the Kubernetes API server is alive (i.e., it has not crashed).

Used by kubelet to determine if the API server needs to be restarted.

# kubectl get --raw /livez

3.General Health Check (/healthz)

Checks the overall health of the API server

 # kubectl get --raw /healthz

4.Verify etcd is Healthy.

# kubectl get --raw /readyz?verbose

 

Deploying kube-state-metrics in Kubernetes.

# helm repo add prometheus-community https://prometheus-community.github.io/helm-charts

# helm repo update

# helm install kube-state-metrics prometheus-community/kube-state-metrics

# kubectl get pods -n default

# kubectl port-forward svc/kube-state-metrics 8080

# curl http://localhost:8080/metrics

                        or

# kubectl  get pods

# kubectl expose pod kube-state-metrics-5495d45756-p89mm --type=NodePort --port=8080 --name=metrics-svc

# kubectl describe svc metrics-svc

# curl http://10.107.59.111:8080/metrics

# curl http://10.107.59.111:8080/metrics | grep kube_node_info

# kubectl run nginx --image=nginx --port=80

# kubectl get pods

# curl http://10.107.59.111:8080/metrics | grep kube_pod_status_phase | grep "nginx

Common Metrics from kube-state-metrics

Metric Name	Description
kube_pod_status_phase	Shows the phase (Pending, Running, Succeeded, Failed) of each pod
kube_node_status_ready	Indicates if a node is ready (1 = Ready, 0 = Not Ready)
kube_deployment_status_replicas	Shows the number of replicas in a deployment
kube_statefulset_replicas	Shows the number of replicas in a StatefulSet

1️.Metrics Collection & Storage Tools

These tools collect real-time metrics and store them for analysis.

Tool	Features
Prometheus	Most popular, collects time-series data, integrates with Grafana & Alertmanager
cAdvisor	Built into Kubelet, provides container-level CPU, memory, and network stats
Kube-State-Metrics	Collects cluster state metrics (Pods, Deployments, Nodes) for Prometheus
Metrics Server	Provides CPU & memory metrics for Horizontal Pod Autoscaler (HPA)
InfluxDB	High-performance time-series database, alternative to Prometheus
OpenTelemetry	Standardized observability framework for traces, metrics, and logs

2️.Monitoring & Visualization Tools

These tools provide dashboards and real-time data visualization.

Tool	Features
Grafana	Best for visualizing Prometheus metrics, customizable dashboards
Kibana	Works with Elasticsearch for logging & metric visualization
Thanos	Extends Prometheus for long-term storage and high availability
Chronograf	Works with InfluxDB, provides dashboards & alerting

3️.Logging & Event Monitoring

These tools focus on log collection, indexing, and analysis.

Tool	Features
Elasticsearch + Kibana (ELK Stack)	Best for searching & analyzing logs
Loki (by Grafana)	Log aggregation, lightweight alternative to ELK
Fluentd	Collects logs and sends them to various backends (ELK, Loki, Splunk)
Logstash	Part of ELK stack, processes & filters logs
Graylog	Centralized log management with alerting features

4️.Distributed Tracing & Performance Monitoring

These tools help track requests across microservices.

Tool	Features
Jaeger	Distributed tracing, tracks requests across services
Zipkin	Similar to Jaeger, collects trace data from microservices
OpenTelemetry	Standardized observability framework (tracing, metrics, logs)

5️.Kubernetes-Native Monitoring & Cloud Solutions

These tools are cloud-native and integrate with Kubernetes.

Tool	Features
Datadog	SaaS-based K8s monitoring & security
New Relic	Full observability (metrics, logs, traces)
Dynatrace	AI-powered monitoring for Kubernetes, cloud, and apps
Google Cloud Operations (Stackdriver)	GCP-native monitoring for GKE
Azure Monitor for Containers	Azure-native monitoring for AKS
Amazon CloudWatch	AWS-native monitoring for EKS

  Prometheus (metrics) + Grafana (visualization) + Loki (logs) + Jaeger (tracing).