🚀 Getting Started
Introduction​
This guide provides step-by-step instructions for deploying Kubesense, an eBPF-based observability tool, in an on-premise Kubernetes environment using Helm. By the end of this guide, you'll have a fully operational deployment of Kubesense to monitor and trace network traffic and resource usage across your Kubernetes clusters.
Overview​
The deployment process for on-premises deployments of Kubesense using Helm can be executed in two different ways: either by deploying all server components within the same cluster or by deploying server and sensor elements in separate clusters.
Prerequisites​
Before proceeding with the deployment, ensure the following requirements are met:
Kubernetes cluster: An operational Kubernetes cluster (version 1.21+ recommended). Helm: Installed and configured (version 3.x recommended). Node Permissions: Adequate permissions to install Helm charts and deploy pods in the Kubernetes cluster. eBPF Support: Nodes in the Kubernetes cluster must support eBPF, which is available on most modern Linux distributions (kernel version 4.18+).
If you haven't installed Helm yet, follow these steps:​
curl https://raw.githubusercontent.com/helm/helm/master/scripts/get-helm-3 | bash
Verify that Helm is installed correctly:​
helm version
Monitor and Explore​
Once Kubesense is up and running, you can explore its rich observability features such as:
- Network Traffic Monitoring: Visualize traffic between pods and services using eBPF.
- Latency Tracking: Measure latency across service communications.
- Resource Usage Monitoring: Get detailed insights into CPU, memory, and disk usage per pod.