[mrbabbage]

Thanks Simon! I can indeed confirm that this script managed to saturate the database's hardware capacity (I recall CPU being the bottleneck, and I had to dial down the parallelism to leave some CPU for actual application queries).