Decoupling Jenkins Controller Storage via OCI Registries with ORAS Artifact Manager

The Jenkins project highlighted the ORAS Artifact Manager Plugin as its featured capability, providing a cloud-native mechanism for offloading archived build artifacts and pipeline stashes to standard OCI-compliant container registries instead of storing them on Jenkins controller local disks. Built on top of the ORAS (OCI Registry As Storage) Java SDK and its corresponding Jenkins API plugin, the tool maps Jenkins jobs into OCI repositories and utilizes OCI 1.1 artifact manifests to anchor individual build outputs to root build tags via standard subject referrers. This development directly tackles one of the most persistent operational pain points in enterprise Jenkins administration: controller disk exhaustion. Storing large build artifacts, test reports, and inter-stage stashes on the controller disk turns CI masters into stateful bottlenecks that are cumbersome to back up, scale, and recover. By redirecting this payload directly to OCI registries—such as Harbor, Amazon ECR, Azure Container Registry, or Google Artifact Registry—platform teams preserve the familiar Jenkins UI artifact experience while delegating storage durability, retention policies, and geo-replication to dedicated registry infrastructure. Architecturally, this reflects the broader cloud-native consolidation around OCI standards. Over recent years, container registries have evolved beyond simply hosting Docker images to serving as generic, standardized artifact repositories for Helm charts, Software Bills of Materials (SBOMs), and provenance attestations. Aligning Jenkins build artifact lifecycles with the OCI 1.1 specification eliminates the need for separate blob storage connectors or specialized artifactory tooling in environments that already maintain centralized container registries. In practice, DevOps teams operating self-hosted or Kubernetes-managed Jenkins clusters should evaluate migrating from local disk artifact archiving to OCI-backed storage. Doing so significantly simplifies moving toward ephemeral controller models where Jenkins masters can restart without risking build data loss. However, teams must monitor the network bandwidth between CI workers, controllers, and registries, and ensure registry access credentials and retention cleanup routines are properly aligned with CI pipeline volume.