Hi aurora-dev, I'd like to propose a feature that I think could be useful to admins/operators of Aurora clusters. The feature would be to allow an administrator of the cluster to "block" write operations from happening in the scheduler, essentially putting it into read-only mode for a period of time. It could be a simple global lock across the entire cluster or we could get sophisticated and apply locks across roles, environments or specific jobs.
I think this has the potential to be very simple (reject new RPCs but allow existing operations to continue) or it can start to get complicated (also pause all in-progress updates and other reschedule events). Would welcome feedback on this idea before I create a ticket. Thanks, David
