An instance lifecycle change should identify the exact target and the intended application impact. A soft stop and a start are control-plane actions, not proof that a business job completed or that its data is recoverable. Preserve the image, subnet and resource settings while carrying out the requested maintenance.
These original worker tasks use supplied OCIDs. One stopped Flex instance also receives a reviewed CPU and memory allocation. The browser stores those states without executing a guest or validating every real shape-change constraint. A production maintenance plan needs the current provider requirements, a safe application handoff and the original workload test afterward. Do not substitute a successful command response for those service and data checks.
Use the exact instance identifier and verify the required lifecycle outcome.