Containers in a pod do not have egress or ingress network restrictions. A compromised container could be exploited for activities such as unauthorised data exfiltration, establishing a reverse shell to maintain access to the cluster, or launching attacks against containers in different namespaces. Kubernetes namespaces, in particular, do not inherently restrict network access. This can allow lateral movement within a cluster.

Remediation

Apply network policy. The following examples restricts all egress traffic from "appspace" namespace.

apiVersion: networking.k8s.io/v1
kind: NetworkPolicy
metadata:
  name: default-deny
  namespace: appspace
spec:
  podSelector: {}
  policyTypes:
  - Egress

---
apiVersion: v1
kind: Namespace
metadata:
  name: appspace