Review policies before enabling work
Inspect renewal and lifecycle settings, grace periods and the actual products affected. Confirm the scheduler and queue are healthy before relying on recurring invoices, reminders or provider actions.
Use explicit policy choices for automatic suspension/restoration or termination. Enabling a scheduler does not authorise every possible provider write; each action still depends on product, connection, payment, ownership and lifecycle checks.
Keep customer intent visible
Customers can review renewals and submit cancellation requests through their permitted account tools. A cancellation request is for staff review; it is not proof that a remote service was immediately terminated.
Review the correct service, effective date, invoice/payment state and provider identity before taking an irreversible action. Retain the relevant history and backups.
Investigate failures
Review task health and exact operation references. A job can be waiting, failed, stale or uncertain for different reasons. A reservation expiring is not proof the remote provider did nothing.
After an uncertain create, renewal, suspension or termination, inspect the original provider state and reconcile it through the supported workflow. Do not clear submission markers, replay a live job as a test or create a replacement account to conceal an unresolved operation.
Rehearse the intended failure and recovery journey on staging before enabling it for customers.