27 Oct 2026
Container and Kubernetes Security Fundamentals for Growing Teams
Kubernetes isn't secure by default, and most teams adopt it faster than the controls it needs. A practical guide to the fundamentals that matter.
Container adoption has moved well past the early-adopter enterprise teams it started with — a huge share of SMB software teams now run at least part of their stack on Kubernetes or a managed equivalent, often because a cloud provider’s managed offering made it the path of least resistance rather than a deliberate architectural choice. What doesn’t scale down automatically with team size is the security configuration Kubernetes assumes you’ll do yourself.
This is a practical, non-enterprise guide to the container and Kubernetes security fundamentals that actually matter — grounded in real, currently tracked vulnerabilities in the Kubernetes ecosystem, not generic “best practices” advice disconnected from what’s actually being exploited.
Why Kubernetes needs deliberate hardening
A fresh Kubernetes cluster is optimized for one thing: getting workloads running with minimal friction. That design choice has real security consequences by default:
- Pods can reach other pods freely unless you define network policies restricting that traffic — there’s no implicit segmentation between workloads in the same cluster.
- Service accounts get broad, automatically mounted tokens unless you explicitly disable or scope auto-mounting, meaning a compromised container can often use its own service account token to query the Kubernetes API far beyond what that specific workload needs.
- RBAC (role-based access control) starts permissive in many setup paths, and cluster-admin bindings created during initial setup or testing have a habit of never getting removed once the cluster goes to production.
None of this is a flaw in Kubernetes — it’s a tool built for flexibility, and flexibility by default means permissive by default. The responsibility to narrow that down sits with whoever operates the cluster.
Real, current vulnerabilities worth knowing about
Beyond configuration risk, Kubernetes-adjacent components carry their own vulnerabilities, tracked through the project’s official CVE feed. As of this writing, that feed lists several that illustrate the kind of risk container-running teams should actually watch for:
- Ingress-nginx configuration injection (tracked as CVE-2026-4342, with a related issue CVE-2025-15566) — a flaw in how the widely used ingress-nginx controller handles comment-based and header-based nginx configuration, which can allow configuration injection if untrusted input reaches it.
- NFS CSI driver path traversal (CVE-2026-3864) — a path traversal issue in the CSI (Container Storage Interface) driver for NFS that can result in deletion of unintended directories on the NFS server backing persistent storage.
The pattern worth noting: these aren’t flaws in Kubernetes’ core scheduler or API server — they’re in the ingress controllers and storage drivers that nearly every real cluster depends on. Keeping Kubernetes itself patched isn’t enough; the ecosystem of controllers and drivers around it needs the same patch discipline. Teams running ingress-nginx or an NFS-backed CSI driver should check the official Kubernetes CVE feed for affected and fixed versions directly before treating either as resolved.
A practical hardening checklist
Access control:
- Review every ClusterRoleBinding and RoleBinding for
cluster-adminor similarly broad roles; remove any left over from setup or testing - Scope RBAC per-namespace and per-workload rather than granting cluster-wide access as a default
- Disable automatic service account token mounting (
automountServiceAccountToken: false) for pods that don’t need to call the Kubernetes API
Network and workload isolation:
- Define network policies that restrict pod-to-pod traffic to what’s actually needed — don’t rely on the cluster’s permissive default
- Run workloads with the most restrictive Pod Security Standard your applications can tolerate (
restrictedwhere possible) - Avoid running containers as root inside the pod unless a specific workload genuinely requires it
Supply chain and image integrity:
- Pull images only from a trusted, access-controlled registry — not directly from public registries in production
- Use signed images and verify signatures as part of your deployment pipeline where your tooling supports it
- Scan images for known vulnerabilities before deployment, and keep a process for re-scanning images already running in the cluster as new CVEs are disclosed
Operational hygiene:
- Subscribe to or periodically check the official Kubernetes CVE feed, especially for ingress controllers, CSI drivers, and any other third-party components your cluster depends on
- Keep Kubernetes itself, plus every add-on component, on a supported, patched version
- Centralize and monitor Kubernetes audit logs rather than relying on default retention
Who should prioritize this
Any team running production workloads on Kubernetes — even a small one — should work through the access-control and network-isolation sections at minimum; both are largely one-time configuration reviews rather than ongoing overhead. Teams using ingress-nginx or NFS-backed persistent storage should specifically verify their versions against the official CVE feed given the vulnerabilities above. Teams still evaluating whether to adopt Kubernetes at all should weigh this operational security overhead honestly against whether a simpler container orchestration approach might fit a small team better.
Key takeaways
Kubernetes is not secure by default — it’s flexible by default, and the hardening work (RBAC scoping, network policies, service account restrictions) is a deliberate step most teams skip under deployment pressure. Real, currently tracked vulnerabilities in widely used components like ingress-nginx and NFS CSI drivers show that the ecosystem around Kubernetes needs the same patch discipline as Kubernetes itself. The highest-impact fixes — removing stale cluster-admin bindings and scoping service account tokens — are configuration reviews a small team can realistically complete, not an enterprise-scale program.
Shahid Security’s Network & Cloud Security service reviews cloud and container infrastructure configuration as part of a broader security assessment, alongside Custom Software Development for teams building on containerized infrastructure from the ground up.
Not sure whether your Kubernetes setup has configuration gaps like these? Book a free consultation.
Sources: Kubernetes official CVE feed, Checkmarx: Kubernetes Security Risks, Technologies, and 9 Best Practices.
FAQ
Is Kubernetes secure out of the box?
No. A default Kubernetes installation prioritizes getting a working cluster running, not locking it down. Role-based access control (RBAC) needs to be explicitly scoped, network policies need to be explicitly written (by default, pods can talk to any other pod in the cluster), and service account tokens are often mounted far more broadly than any given workload actually needs. Hardening is a deliberate step after installation, not something that happens automatically.
My team is small — do we really need dedicated Kubernetes security practices?
If you're running production workloads on Kubernetes, yes, though it doesn't need to be a big program. The highest-impact steps (reviewing RBAC bindings, scoping service account tokens, using signed images from a trusted registry) are one-time or low-maintenance changes, not an ongoing operational burden. The risk isn't proportional to team size — a misconfigured cluster is just as exploitable whether five people or five hundred manage it.
What is the most common real-world cause of Kubernetes breaches?
Overly broad access, not an exotic exploit. Two patterns show up repeatedly in post-incident analysis: service accounts with far more permissions than the workload needs (so a single compromised pod can pivot broadly), and unused cluster-admin role bindings left over from initial setup or testing that nobody removed. Both are configuration problems, not vulnerabilities in Kubernetes itself, which is exactly why a periodic RBAC and access review is one of the highest-value things a team can do.

Written by Shahid Iqbal
Founder & Cybersecurity Consultant, Shahid Security
