Deploy in your cluster
Platform

Runtime security built for how Kubernetes actually works.

eBPF probes at the kernel level. DaemonSet deployment. Falco-compatible rules. A policy engine that evaluates events in microseconds.

eBPF Engine

Kernel-level visibility without kernel modules.

Before eBPF, runtime security required fragile kernel modules or sidecar agents with measurable latency overhead. eBPF lets us attach probes directly to the syscall table — no restarts, no application changes, no performance penalty.

Zero-copy ring buffer

Syscall events flow from kernel to userspace via a shared memory ring buffer — no data copy, minimal CPU overhead. Handles 4,000+ events/second per node.

Kernel safety guarantee

eBPF programs are verified by the kernel verifier before loading — they cannot crash or corrupt kernel memory, unlike loadable kernel modules.

Hot-reload rule updates

Push policy changes without restarting the DaemonSet or any application pod. New rules take effect in under 5 seconds across all nodes.

Policy Engine

YAML policies. Falco-compatible syntax.

Write detection rules in a declarative YAML format that's compatible with Falco's rule language. Import your existing OSS Falco rules without modification.

detection-policies/container-escape.yaml Falco-compatible
- rule: container_escape_via_namespace
  desc: Detect container namespace escape via unshare syscall
  condition: syscall.type = unshare and container.id != host
            and proc.args contains CLONE_NEWPID
  output: "Container escape attempt (pod=%k8s.pod.name ns=%k8s.ns.name pid=%proc.pid)"
  priority: CRITICAL
  tags: [container, escape, cis_kubernetes]
 
# Kubesentry extensions: alert routing + action
  actions:
    - type: alert
      channels: [slack:#sec-ops, pagerduty:p1-on-call]
    - type: isolate_pod
      confirm: false # auto-execute

Detection coverage matrix.

Threat Category Syscall Signals Audit Log Network Baseline Drift Falco Rule
Container Escape
Cryptomining
Lateral Movement
Privilege Escalation
Syscall Anomaly

Deployment options.

Helm Chart

One command install via our official Helm chart. Supports values.yaml customization for resource limits, namespace targets, and alert routing.

$ helm install ks kubesentry/kubesentry

Kubernetes Operator

Deploy the Kubesentry Operator for GitOps-native management. Reconciles DaemonSet state against your KubesentryConfig CRD automatically.

Manual YAML

Raw Kubernetes manifests available for air-gapped or restricted environments. No external registry dependencies when using the bundle.

Ready to deploy?

Takes under 10 minutes on any CNCF-conformant K8s distribution.