Problem
Application teams had to build and release software inside a high-security, air-gapped data centre. Connected environments condition us to pull images from public registries and iterate rapidly. At an air-gapped site none of that holds: every package, repository and image comes through the mirror, and every release has to be secure before it reaches the cluster.
The team was small, so the platform had to let it do more with less.
Constraints
- No internet. The registry server supplies the packages, Git repositories and container images a developer needs to build and release software.
- Security at every stage. An insecure image must never reach the cluster, and the cluster has to be watched at runtime.
- A small team. Automation has to carry the repetitive work, so the team doesn’t rely on specialists for every check.
- Releases are tested first. Every change goes through System Integration and User Acceptance Testing before production.
Decision
Start from the local registry. The registry server and a private Git server are the foundation: everything a developer builds from comes through them. Our custom container workloads are built on Red Hat’s Universal Base Images, which are hardened and receive regular vulnerability patches and security advisories.
Workspaces defined as code. OpenShift Dev Spaces gives developers a cloud-native development environment in a browser. A Devfile defines each workspace as code, with every container and resource it needs, so every workspace runs from approved images with a consistent configuration.
Quality gates in the pipeline. A commit to the Quality Assurance branch triggers validation streams in an OpenShift Pipelines (Tekton) pipeline. They enforce unit tests, code coverage and coding standards, and run in parallel for a shorter feedback loop.
Scan before promotion. A Red Hat Advanced Cluster Security scan in the promotion pipeline identifies known vulnerabilities and flags insecure configurations, such as running containers as root. If it finds critical issues, promotion is blocked and developers are notified.
Release through GitOps. Once validated, the image manifest is committed to the configuration repository, and Argo CD rolls the new image out to the cluster for System Integration and User Acceptance Testing.
Detect at runtime. Advanced Cluster Security monitors the cluster for anomalous processes, suspicious network activity and privilege escalations, and blocks policy violations as they happen.
Outcome
- 4
- operators carry the workflow from commit to release
- Minutes
- for a developer to be productive, not days
- Security. Every image is built on a secure foundation, and the registry holds only verified, trusted artefacts.
- Speed. A developer can be productive in minutes, not days, which lowers onboarding cost.
- Continuous assurance. Automated pipelines enforce testing, building and scanning, so by release the software is already proven.
- Confidence. Because the cluster is managed by GitOps, every deployment is consistent and automatic.
For your team
Secure by design means every stage, from a hardened base image to scanning before promotion and detection at runtime. Automation is what lets a small team do it.
My role
I designed the internal developer platform and led the team that built it, and presented it at IBM TechXchange 2025 in Orlando.