Kubernetes Monitoring Part1

Kubernetes Monitoring Tools: Enhancing System Health and Performance

Kubernetes, while powerful, introduces complexity. Its ability to distribute containerized applications across data centres and cloud providers necessitates robust monitoring solutions. Let’s explore key open-source tools for Kubernetes monitoring:

1. Prometheus:

• Developed by SoundCloud and donated to CNCF, Prometheus excels at monitoring Kubernetes and Docker.
• Provides actionable metrics and alerts.
• Often paired with Grafana for visualization.

 

2. Grafana:

• An open-source analytics and metric visualization platform.
• Offers dashboards for Cluster, Node, Pod/Container, and Deployment monitoring.
• Integrates seamlessly with Prometheus.

 

3. Jaeger:

• A tracing system for complex distributed systems.
• Troubleshoots transactions, addressing issues like latency optimization.

 

4. Kubernetes Dashboard:

• Web UI add-on for workload health monitoring.

 

5. Prometheus and Grafana Integration:

• Prometheus, a CNCF standard, stands out among time-series databases.
• Grafana enhances Prometheus with customizable dashboards.
• Grafana’s external visualization complements Prometheus.
• When deploying Prometheus, Grafana is often included by default.
• Detailed Grafana dashboards allow namespace and pod-level monitoring.

 

6. Deployment Strategy:

• Deploy Prometheus and Grafana together.
• Consider additional metrics collection tools like Alert-manager, Metrics server, and Kube-metrics.
• These tools provide comprehensive insights into your Kubernetes cluster.

Streamlining Kubernetes Monitoring with EFS and Persistent Volumes

In our deployment strategy, we leverage an existing Elastic File System (EFS)  for monitoring purposes. Rather than creating new mount points, we utilize the same ones established for mounting WordPress and Jenkins. Our approach involves creating dedicated directories for configuration files and data within the existing EFS setup.

Key Considerations:

1. EFS Integration: We assume that an EFS has already been set up, as detailed in the provided blog.
2. Reuse of Mount Points: By reusing existing mount points, we maintain consistency and avoid unnecessary complexity.
3. Persistent Volumes (PVs): We rely on PVs to manage storage resources for monitoring components.

 

Prerequisites for Effective Monitoring:

• Ensure that Kubernetes pods are actively running within your cluster.
• Utilize both Master and Worker nodes to monitor hardware resources effectively.

 

By following these guidelines, we can establish efficient monitoring practices while maximizing the benefits of our existing infrastructure.

Step1: Setting Up Monitoring Infrastructure for Kubernetes

1. Create Directories on the Master and Worker Nodes:

1. Log in to the Master Node as the “jenkins” user.
2. Navigate to the home directory: $ cd /home/jenkins.
3. Create a directory named “monitoring”: $ mkdir monitoring.

 

2. Namespace Configuration:

• We’ll use the “monitoring” Namespace for most of our monitoring deployments.
• Additionally, the default “kube-system” Namespace collects metrics.
• Create a YAML file for the “monitoring” Namespace (e.g., monitoring-namespace.yaml).
• Define the Namespace:

apiVersion: v1
kind: Namespace
metadata:
  name: monitoring

• Apply the configuration: $ kubectl create -f monitoring-namespace.yaml.

 

3. Download Monitoring Files:

• Within the “monitoring” directory, clone the necessary files from the repository:

$ git clone https://github.com/AvangelsTech/monitoring.git .

• This command clones all relevant directories into the current “monitoring” directory.

 

By following these steps, you’ll establish a solid foundation for monitoring your Kubernetes cluster.

In this step, we’ll set up the necessary metrics for Prometheus integration. Follow these organized instructions:

1. Navigate to the Monitoring Directory:

• Log in to the Master Node as the “jenkins” user.
• Change to the monitoring directory: $ cd monitoring.
• Confirm that you have the required directories and YAML files.

 

2. Cloning Monitoring Files:

• We’ve already cloned the GitHub repository containing monitoring files.
• Verify the directory structure: $ ls -ltr.
• You should see subdirectories like alert-manager, grafana, kube-metrics, metrics-server, and prometheus.

drwxrwxr-x. 2 jenkins jenkins 6144 Apr 17 07:00 alert-manager
drwxrwxr-x. 2 jenkins jenkins 6144 Apr 16 10:23 grafana
drwxrwxr-x. 2 jenkins jenkins 6144 Jan7 14:04 kube-metrics
drwxrwxr-x. 2 jenkins jenkins 6144 Jan 10 08:57 metrics-server
drwxrwxr-x. 2 jenkins jenkins 6144 Apr 17 06:00 prometheus

 

3. Deploy Alert-manager:

• Execute: $ kubectl create -f alert-manager/.
• This command sets up the alert-manager pod, services, and deployment within the “monitoring” Namespace.
• Note: The alert manager is exposed on port 31000.

 

4. Access Alert-manager:

• Use the following syntax in a web browser:

http://<ip-of-worker-node>:31000

• You’ll be able to view the alert-manager interface.

Remember to replace <ip-of-worker-node> with the actual IP address of your worker node.

Deploying “kube-metrics” and “metrics-server” for Kubernetes Monitoring

1. Deploy “kube-metrics”:

• Execute the following command:

$ kubectl create -f kube-metrics/

• This action sets up the “kube-metrics” pod, services, and deployment within the “kube-system” Namespace.

 

2. Deploy “metrics-server”:

• Run the command:

$ kubectl create -f metrics-server/

• This creates the “metrics-server” components in the same “kube-system” Namespace.

 

By following these steps, you’ll have essential monitoring tools in place for your Kubernetes cluster.