Deployment Context
Before deploying KubeSense, understand how its two core components — the server and sensor — work together to provide observability.
Components
Sensor
The sensor is deployed as a DaemonSet on each node in your Kubernetes cluster. It uses eBPF to collect telemetry data directly from the kernel:
- Traces — Network requests between pods and services
- Logs — Container and application log streams
- Metrics — CPU, memory, disk, and network utilization
- Latency — End-to-end request timing across services
Sensors are lightweight and operate with minimal overhead. Collected data is forwarded to the server for processing.
Server
The server receives data from one or more sensors, processes it, and makes it available through the KubeSense dashboard:
- Data aggregation from multiple clusters and sensors
- Storage in optimized datastores for logs, traces, and metrics
- Dashboard for visualization, querying, and alerting
Deployment Models
| Model | Description |
|---|---|
| In-Cluster | Server and sensor deployed in the same cluster. Simplest setup for single-cluster environments. |
| Multi-Cluster | Server deployed centrally, sensors deployed in each cluster to monitor. Ideal for organizations with multiple clusters. |
Pre-Deployment Checklist
Sensor requirements:
- Kubernetes nodes must support eBPF (Linux kernel 4.18+)
- Network connectivity to the server endpoints (see Networking)
Server requirements:
- Persistent storage for data retention
- External hostname or Ingress for dashboard access
Next Steps
Deploy Server and Sensor Together (In-cluster)
Deploy Server and Sensor Separately (Multi-cluster)