Skip to main content

Kubernetes Monitoring Stack

Introduction​

When managing Kubernetes clusters, having visibility into what's happening inside is crucial. A monitoring stack allows you to collect metrics, visualize data, and set up alerts for your Kubernetes infrastructure. This guide will walk you through building a comprehensive monitoring solution for your Kubernetes clusters using popular open-source tools.

Think of a monitoring stack as the dashboard of a car - it shows you important information about what's happening "under the hood" of your Kubernetes environment, helping you detect issues before they become problems.

Why Monitoring Matters in Kubernetes​

Kubernetes environments are dynamic and complex. Without proper monitoring:

  • You won't know if your applications are healthy
  • Resource bottlenecks may go undetected
  • Troubleshooting becomes difficult
  • Scaling decisions lack data

Core Components of a Kubernetes Monitoring Stack​

A typical Kubernetes monitoring stack consists of these primary components:

Let's break down each component and how they work together.

Setting Up Prometheus for Metrics Collection​

Prometheus is the de-facto standard for Kubernetes metrics collection.

Installing Prometheus with Helm​

Helm makes it easy to install Prometheus on your Kubernetes cluster:

bash
# Add the Prometheus community Helm repository
helm repo add prometheus-community https://prometheus-community.github.io/helm-charts
helm repo update

# Install Prometheus
helm install prometheus prometheus-community/prometheus -n monitoring --create-namespace

Understanding Prometheus Configuration​

Prometheus uses a YAML configuration file. Here's a simplified example:

yaml
global:
  scrape_interval: 15s

scrape_configs:
  - job_name: 'kubernetes-pods'
    kubernetes_sd_configs:
      - role: pod
    relabel_configs:
      - source_labels: [__meta_kubernetes_pod_annotation_prometheus_io_scrape]
        action: keep
        regex: true

How Prometheus Works​

Prometheus works by "scraping" metrics from HTTP endpoints exposed by your applications and Kubernetes components. It stores these metrics in a time-series database.

To make your application scrape-able by Prometheus, you need to:

  1. Expose metrics on an HTTP endpoint (commonly /metrics)
  2. Format the metrics in the Prometheus format
  3. Add the right annotations to your Kubernetes resources

Here's an example of a Pod with Prometheus annotations:

yaml
apiVersion: v1
kind: Pod
metadata:
  name: example-app
  annotations:
    prometheus.io/scrape: "true"
    prometheus.io/port: "8080"
    prometheus.io/path: "/metrics"
spec:
  containers:
    - name: example-app
      image: example-app:latest
      ports:
        - containerPort: 8080

Setting Up Grafana for Visualization​

Grafana turns your metrics into visual dashboards.

Installing Grafana​

bash
# Add Grafana Helm repository
helm repo add grafana https://grafana.github.io/helm-charts
helm repo update

# Install Grafana
helm install grafana grafana/grafana -n monitoring

Getting the Grafana Password​

bash
kubectl get secret --namespace monitoring grafana -o jsonpath="{.data.admin-password}" | base64 --decode ; echo

Connecting Grafana to Prometheus​

Once installed, you'll need to:

  1. Log in to Grafana UI
  2. Go to Configuration > Data Sources
  3. Add a new Prometheus data source with URL: http://prometheus-server.monitoring.svc.cluster.local

Creating Your First Dashboard​

Let's create a simple dashboard to monitor CPU and memory usage:

  1. In Grafana, click "Create Dashboard"
  2. Add a new panel
  3. Use this PromQL query for CPU usage:
sum(rate(container_cpu_usage_seconds_total{pod=~"$pod"}[5m])) by (pod)
  1. Add another panel with this query for memory usage:
sum(container_memory_usage_bytes{pod=~"$pod"}) by (pod)

Adding Alerting with AlertManager​

AlertManager handles alerts sent by Prometheus and routes them to the right receiver (email, Slack, etc.).

Configuring AlertManager​

Create an alertmanager.yaml configuration file:

yaml
global:
  resolve_timeout: 5m

route:
  group_by: ['alertname', 'job']
  group_wait: 30s
  group_interval: 5m
  repeat_interval: 12h
  receiver: 'email-notifications'

receivers:
- name: 'email-notifications'
  email_configs:
  - to: '[email protected]'
    from: '[email protected]'
    smarthost: 'smtp.example.com:587'
    auth_username: 'alertmanager'
    auth_password: 'password'

Creating Alert Rules in Prometheus​

Alert rules tell Prometheus when to fire alerts. Here's an example rule:

yaml
groups:
- name: example
  rules:
  - alert: HighCPUUsage
    expr: sum(rate(container_cpu_usage_seconds_total{namespace!="kube-system"}[5m])) by (pod) > 0.8
    for: 5m
    labels:
      severity: warning
    annotations:
      summary: "High CPU usage detected"
      description: "Pod {{ $labels.pod }} has high CPU usage for more than 5 minutes."

Collecting and Analyzing Logs with the EFK Stack​

While metrics show you what is happening, logs tell you why. The EFK stack (Elasticsearch, Fluentd, Kibana) is commonly used for log management in Kubernetes.

Installing Fluentd​

bash
# Create a namespace for logging
kubectl create namespace logging

# Apply Fluentd configuration
kubectl apply -f - <<EOF
apiVersion: v1
kind: ServiceAccount
metadata:
  name: fluentd
  namespace: logging
---
apiVersion: rbac.authorization.k8s.io/v1
kind: ClusterRole
metadata:
  name: fluentd
rules:
- apiGroups:
  - ""
  resources:
  - pods
  - namespaces
  verbs:
  - get
  - list
  - watch
---
apiVersion: rbac.authorization.k8s.io/v1
kind: ClusterRoleBinding
metadata:
  name: fluentd
roleRef:
  apiGroup: rbac.authorization.k8s.io
  kind: ClusterRole
  name: fluentd
subjects:
- kind: ServiceAccount
  name: fluentd
  namespace: logging
EOF

Installing Elasticsearch and Kibana​

bash
# Add Elastic Helm repository
helm repo add elastic https://helm.elastic.co
helm repo update

# Install Elasticsearch
helm install elasticsearch elastic/elasticsearch -n logging

# Install Kibana
helm install kibana elastic/kibana -n logging

Integrating kube-state-metrics​

To get additional metrics about the state of Kubernetes objects, you can use kube-state-metrics:

bash
# Install kube-state-metrics
helm install kube-state-metrics prometheus-community/kube-state-metrics -n monitoring

This provides metrics like:

  • Pod status (running, pending, failed)
  • Deployment status (desired vs available replicas)
  • Node capacity and allocatable resources

Creating a Custom All-in-One Monitoring Stack​

For beginners, setting up each component individually can be complex. Let's create a simple all-in-one setup:

bash
# Create a values.yaml file for the Prometheus Stack
cat > monitoring-values.yaml <<EOF
grafana:
  enabled: true
  adminPassword: admin-password
  
prometheus:
  prometheusSpec:
    retention: 15d
    resources:
      requests:
        memory: 512Mi
        cpu: 500m
      limits:
        memory: 2Gi
        cpu: 2000m

alertmanager:
  enabled: true
  config:
    global:
      resolve_timeout: 5m
    route:
      group_by: ['alertname', 'job']
      group_wait: 30s
      group_interval: 5m
      repeat_interval: 12h
      receiver: 'null'
    receivers:
    - name: 'null'

nodeExporter:
  enabled: true

kube-state-metrics:
  enabled: true
EOF

# Install the Prometheus Stack with our custom values
helm install monitoring prometheus-community/kube-prometheus-stack \
  -f monitoring-values.yaml \
  -n monitoring \
  --create-namespace

Practical Exercise: Monitor a Sample Application​

Let's put everything together by deploying and monitoring a sample application:

  1. Deploy a sample application:
bash
kubectl apply -f - <<EOF
apiVersion: apps/v1
kind: Deployment
metadata:
  name: sample-app
  labels:
    app: sample-app
spec:
  replicas: 3
  selector:
    matchLabels:
      app: sample-app
  template:
    metadata:
      labels:
        app: sample-app
      annotations:
        prometheus.io/scrape: "true"
        prometheus.io/port: "8080"
    spec:
      containers:
      - name: sample-app
        image: nginx:latest
        ports:
        - containerPort: 80
---
apiVersion: v1
kind: Service
metadata:
  name: sample-app
spec:
  selector:
    app: sample-app
  ports:
  - port: 80
    targetPort: 80
EOF
  1. Port-forward to access Grafana:
bash
kubectl port-forward svc/monitoring-grafana 3000:80 -n monitoring

  1. Import a Kubernetes dashboard in Grafana (Dashboard ID: 10000)

  2. Create an alert if the sample app's pod count drops below 3:

yaml
groups:
- name: sample-app
  rules:
  - alert: SampleAppPodCount
    expr: count(kube_pod_info{namespace="default", pod=~"sample-app.*"}) < 3
    for: 1m
    labels:
      severity: critical
    annotations:
      summary: "Sample App pod count is below desired"
      description: "The number of running pods for the Sample App is below 3."

Best Practices for Kubernetes Monitoring​

  1. Monitor the Four Golden Signals:

    • Latency
    • Traffic
    • Errors
    • Saturation

  2. Implement Proper Resource Requests and Limits: This helps the monitoring system identify when containers are approaching their limits.

  3. Use Labels Consistently: Labels like app, environment, and team make it easier to filter and group metrics.

  4. Set Up Relevant Alerts: Only alert on actionable issues to avoid alert fatigue.

  5. Keep History in Mind: Configure appropriate retention periods for metrics and logs.

Troubleshooting Common Issues​

"No Data Points" in Grafana​

  • Check if Prometheus is correctly scraping metrics
  • Verify the PromQL query syntax
  • Ensure time ranges are set correctly

High Cardinality Issues​

  • Avoid using high-cardinality labels
  • Use recording rules for common queries

Memory Usage Problems in Prometheus​

  • Increase memory limits
  • Reduce retention period
  • Limit the number of time series with better relabeling

Summary​

Setting up a monitoring stack for Kubernetes is essential for maintaining a healthy and efficient cluster. In this guide, we've covered:

  • Setting up Prometheus for metrics collection
  • Visualizing data with Grafana
  • Configuring alerts with AlertManager
  • Collecting logs with the EFK stack
  • Integrating kube-state-metrics for detailed Kubernetes object metrics
  • Creating an all-in-one monitoring solution

By implementing this monitoring stack, you'll gain visibility into your Kubernetes environment, helping you identify and resolve issues before they affect your applications.

Additional Resources​

Exercises for Practice​

  1. Deploy the monitoring stack on a local Kubernetes cluster (like Minikube or Docker Desktop).
  2. Create a custom dashboard in Grafana for a specific application.
  3. Set up an alert that notifies you when pod restarts exceed a threshold.
  4. Configure log collection for a specific namespace.
  5. Create a recording rule in Prometheus to optimize a frequently used query.

If you spot any mistakes on this website, please let me know at [email protected]. I’d greatly appreciate your feedback! :)