Overview
Stakeholder intent for the initial two-location rollout:
- Operate the platform in a single-active mode until a controlled cutover.
- Establish a static intent baseline in Location A with a restricted network OpenShift AI cluster to accelerate DevOps (see ADR-011), using a single ACI pod for this AI use case only.
- Build the dynamic target platform in Location B before switching primary operations, including EVPN/VXLAN multi-tenant networking, multi-cluster platform automation, and a tiered Red Hat Quay registry on the external-profile Clusters as a Service cluster.
- Phase 1 runs production in Location A; Location B is used for platform build, validation, and non-production workloads until cutover.
- Replace the legacy "Platform Services" cluster naming with Clusters as a Service as the target-state shared-services role under Advanced Cluster Management (ACM) governance.
- In the current transitional rollout, run shared services (NetBox, OpenBao, Keycloak, midPoint, NetScaler BLX, GitLab, Red Hat Quay) in Location B Clusters as a Service clusters only.
- Keep Location A as an ACI-scoped OpenShift AI production lane with standalone lifecycle and lane-specific dependency exceptions; Datacenter Services capabilities are not active there.
- Manage server hardware lifecycle via HPE OneView (see Firmware).
- Use Security Operations Center for platform observability (see Security Operations Center).
- Maintain separation between production and development environments through dedicated clusters or tenants and access controls.
Context
These statements capture the accepted rollout baseline and underpin the rollout and cluster naming decisions in the ADRs below.
Applicable Principles
- Reliability and Resilience — rollout protects availability targets.
- Automation First — rollout stages are automated and validated.
- Security by Design — security baselines are enforced.
Decisions
- adr-004-single-active-phased-rollout
- adr-019-hybrid-network-rollout-aci-a-ai-only-evpn-b-target
- adr-005-datacenter-services-baseline
- adr-006-external-load-balancing-procurement
- adr-011-connected-vs-disconnected
- ADR-013 Implementation phasing
Current State
Intent is implemented through explicit controls:
- Cutover governance and release evidence are enforced by adr-004-single-active-phased-rollout and delivery workpackage acceptance criteria.
- Location A AI lane boundary controls: topic-network-aci-location-a-ai-fasttrack and openshift-ai-cluster-setup.
Future Work
- Keep cutover evidence criteria synchronized with runtime telemetry and validation pipelines.