Before pushing a UCS domain profile update to production GPU racks, what should the city Intersight team do?

Before pushing a UCS domain profile update to production GPU racks, the city Intersight team must evaluate impact and follow change control. Domain profile changes can alter FI, network, and domain-wide behavior that running AI jobs depend on, so blast radius is domain-scoped rather than single-blade. Domain 3.2.a treats controlled deployment—impact analysis, maintenance window, rollback plan—as part of using domain profiles in production municipal environments. Applying the updated VLAN and port settings through the GPU server profile templates in Intersight is the wrong mechanism for a domain profile change; domain profiles are domain-scoped and require change control, not server profile template edits. Validating host GPU driver versions against the Intersight compatibility list is a host-level check and does not replace evaluating the domain profile change or following change control before deployment. Backing up running AI training checkpoints and then deploying the updated domain profile to production GPU racks during maintenance still skips the required impact analysis and change-control review before touching production. Exam caveat: some profile elements may be non-disruptive while others are disruptive; the exam answer is still evaluate-then-change-control, not a template edit, driver check, or maintenance-window deploy without prior review. Operational check: diff the proposed domain profile against production, identify disruptive settings, schedule a change window for the affected GPU racks, deploy to a pilot domain first when possible, validate training/inference health, then proceed or roll back per the municipal change record.