Prerequisites
Before deploying K8s Sentinel, confirm your environment meets all of the following requirements.
Environment Requirements
| Component | Minimum Version | Notes |
|---|---|---|
| Kubernetes | 1.32+ | Cluster must be running and accessible via kubeconfig |
| Cilium | 1.14+ | The cluster CNI; kube-proxy replacement, Hubble and Hubble Relay must be enabled (see Installing and Configuring Cilium) |
| Tetragon | 1.0+ | Deployed as a DaemonSet, with gRPC bound to the pod network (0.0.0.0:54321, see Installing Tetragon) |
| Trivy Operator | Optional | Provides the workload vulnerability scan results (v0.60+). The Job and script installations install it into sentinel-system as well; without it the rest of K8s Sentinel works as usual (see Installing Trivy Operator) |
| kubectl | 1.32+ | For local cluster operations; a valid kubeconfig must be configured |
| Access | cluster-admin or TracingPolicy RBAC | Sufficient cluster permissions required for installing K8s Sentinel and operating TracingPolicy CRDs |
Verify Cilium Installation
Run the following command to confirm the Cilium DaemonSet is running normally:
kubectl get pods -n kube-system -l k8s-app=cilium
All Pods should display Running status with a fully ready READY column (e.g., 1/1).
Why: Cilium is deployed as a DaemonSet on every node in the cluster and is responsible for loading eBPF programs into the Linux kernel. Confirming that Pods are READY means the eBPF programs on the corresponding node have loaded successfully - a prerequisite for Tetragon to function.
Verify Tetragon Installation
Run the following command to confirm the Tetragon DaemonSet is running (the namespace may vary depending on installation method):
kubectl get pods -n kube-system -l app.kubernetes.io/name=tetragon
# or
kubectl get pods -n tetragon -l app.kubernetes.io/name=tetragon
Each node should have a corresponding Tetragon Pod in Running state.
Why: The Tetragon DaemonSet deploys one Agent per node, responsible for capturing kprobe security events (syscalls, network connections, file access) via eBPF hooks and forwarding event data to the K8s Sentinel backend for aggregation.
Verify kubectl Connectivity
Confirm your local kubectl is correctly configured and can communicate with the target cluster:
kubectl cluster-info
kubectl get nodes
kubectl cluster-info should display the Kubernetes control plane endpoint address; kubectl get nodes should list all cluster nodes with Ready status.
Verify TracingPolicy CRDs
After Tetragon is installed, the TracingPolicy CRDs are automatically registered in the cluster. Verify they exist:
kubectl get crd | grep tetragon
The output should include the following two CRD resources:
tracingpolicies.cilium.iotracingpoliciesnamespaced.cilium.io
If these CRDs are missing, Tetragon has not completed installation or the CRDs were not applied correctly. Complete the Tetragon installation before continuing.
If Cilium or Tetragon is not yet installed in your environment, follow Installing and Configuring Cilium and Installing Tetragon in this documentation.