Monitoring your cluster using Prometheus and Grafana
Overview¶
This user journey guides you to set up a monitoring facility for the Accuknox control plane cluster and its workloads.
Step 1: Create a NodePool¶
- Create a node pool on GKE / EKS / AKS (or) on-premises worker nodes with below Requirements
| Component Name | Monitoring |
|---|---|
| No of Nodes | 3 |
| Machine Type | E2 / ec2 / vm (vmware / Cloud based VM) |
| Image Type | Any Linux based OS with Container support ( EG: GCP COS-Containerd) |
| CPU Per Node | 6 |
| Memory Per Node | 10 |
| Disk Size Per Node | 50 |
| Total CPU | 18 |
| Total Memory | 30 |
| Total Disk Size | 150 |
| Taints & Lables | monitoring:true |
| Node Pool Name | monitoring |
Nodeselector & Toleration¶
Tolerations:
- key:”monitoring”
operator: “Equal”
value: “true”
effect: “NoSchedule”
Nodeselector:
monitoring: “true”
Install Helm:
- Helm binary is required to proceed further. To install Click here..
Step 2: Install Prometheus and Grafana¶
- Add accuknox repository to install Prometheus & Grafana helm package:
helm repo add accuknox-onprem-monitoring https://USERNAME:[email protected]/repository/accuknox-onprem-monitoring
helm repo update
helm search repo accuknox-onprem-monitoring
helm pull accuknox-onprem-monitoring/grafana-prometheus-stack --untar
kubectl create namespace accuknox-monitoring
kubectl config set-context --current --namespace=accuknox-monitoring
helm install prometheusmetrics grafana-prometheus-stack
kubectl get pods -n accuknox-monitoring
Step 3: Grafana Login¶
Note:
Username: admin
Password: prom-operator
To Change Password:
- Scroll Down Below → Click on Admin Icon → Change Password
Step 4: Grafana Dashboards¶
- Now Click on Dashboards Icon → Browse → Select the Dashboard
List of Dashboards:
