A PersistentVolumeClaim is a Kubernetes request for storage with defined size and class requirements. It allows an application to bind to durable storage without hard-coding the underlying disk details, which helps stateful services obtain the right volume while keeping the deployment portable and repeatable.
Expanded Definition
In Kubernetes, a PersistentVolumeClaim is the request object that lets a workload ask for storage with specific capacity, access mode, and storage class characteristics. It sits between the application and the underlying volume so teams can decouple deployment manifests from the physical or cloud-backed storage implementation.
For NHI and agentic AI workloads, the practical value is governance, not just portability. A PVC can support stateful services that must preserve checkpoints, model artefacts, logs, or queue state while still allowing the underlying storage to be provisioned, replaced, or migrated without changing the application contract. That distinction matters because storage choice often determines durability, encryption posture, backup scope, and who can access the data at rest. Definitions vary across vendors and platform teams on how much policy should be embedded in the storage class versus the claim itself, so the operational model should be documented rather than assumed. The most common misapplication is treating a PVC as a security boundary, which occurs when teams assume the claim alone enforces confidentiality or isolation without verifying the backing volume policy.
For the underlying control expectations around persistent data protection, NIST SP 800-53 Rev 5 Security and Privacy Controls is the most relevant external reference.
Examples and Use Cases
Implementing PersistentVolumeClaims rigorously often introduces storage abstraction overhead, requiring organisations to balance application portability against tighter review of how data is provisioned and retained.
- A stateful database pod requests a PVC so it can restart without losing transaction data, while the storage team controls encryption and backup policy separately.
- An AI inference service mounts a PVC for cached embeddings or model artefacts, allowing the application to stay portable across clusters.
- A CI pipeline writes build outputs to a PVC so repeated jobs can reuse cached dependencies without hard-coding a node-local disk path.
- A secrets-adjacent service stores certificate bundles or token material on persistent storage, which demands careful access review because the PVC does not itself grant secrecy.
For deeper background on how storage misuse can intersect with sensitive data exposure, the DeepSeek breach analysis shows how exposed data paths can amplify downstream risk, and NIST SP 800-53 Rev 5 Security and Privacy Controls remains the best-fit control baseline for storage-related governance.
Why It Matters in NHI Security
Persistent storage often becomes part of the blast radius when an NHI or agentic workload is compromised. If a service account, token, or API key is written into files on a mounted volume, the issue is no longer limited to the process memory of one pod. The data can persist across redeployments, be copied into snapshots, and remain accessible long after the originating workload is gone. NHIMG research on secrets in application security shows that organisations maintain an average of 6 distinct secrets manager instances, a fragmentation pattern that weakens centralised oversight and increases the chance that sensitive material lands in the wrong place.
This is why PVCs matter to NHI governance even though they are not identity objects. They shape where secrets, logs, and operational artefacts live, who can mount them, and how easily they can be recovered after compromise. When storage is misclassified as harmless application plumbing, teams often miss the point where credential sprawl and data persistence combine into an incident. Organisations typically encounter the operational impact only after a leak, pod escape, or incident response review, at which point PersistentVolumeClaim usage becomes unavoidable to investigate and harden.
Standards & Framework Alignment
This section maps relevant standards and security frameworks to the operational risks and controls described in this guidance.
OWASP Non-Human Identity Top 10 address the attack and risk surface, while NIST CSF 2.0, NIST Zero Trust (SP 800-207), NIST SP 800-63 and NIST AI RMF set the governance and control requirements practitioners need to meet.
| Framework | Control / Reference | Relevance |
|---|---|---|
| OWASP Non-Human Identity Top 10 | NHI-02 | Covers improper secret storage risks that can persist on mounted application volumes. |
| NIST CSF 2.0 | PR.DS-1 | Protects data at rest, which includes information persisted on Kubernetes volumes. |
| NIST Zero Trust (SP 800-207) | SC-7 | Zero trust limits implicit trust in storage paths and the workloads that mount them. |
| NIST SP 800-63 | Persistent volumes may store secrets that depend on strong identity assurance around access. | |
| NIST AI RMF | AI systems must manage data lifecycle risks for persistent artifacts and cached outputs. |
Review PVC-mounted paths for secrets exposure and enforce storage handling controls on every workload.
Deepen Your Knowledge
Reviewed and updated by the NHIMG editorial team on September 1, 2026.
NHI Mgmt Group — the #1 independent authority on Non-Human Identity, IAM, and Agentic AI security. nhimg.org