What it means
The platform is highly programmable and depends on identities, admission policy, workload configuration, network policy, supply-chain controls, and runtime monitoring working together.
Why it matters
A small permission or configuration error can create cluster-wide impact because orchestration systems have authority to schedule and change many workloads.
Practical focus
- Restrict cluster and workload privileges
- Use admission and network policies
- Protect the control plane, node identities, and secrets
Common mistakes
- Running containers as privileged by default
- Giving application teams cluster-admin access
- Assuming managed Kubernetes removes customer responsibility
Certification relevance
This subject appears in or supports the following certification bodies of knowledge: