Use `uuid.EnableRandPool()` setting

[colmsnowplow]

I have noticed in investigating another issue that UUID distribution is slightly skewed when using generating UUIDs without setting uuid.EnableRandPool().

This has resulted in a noticeable effect on the distribution of data when setting partition keys - I'm not sure of the specifics of risk of conflicts for ID generation, but I think we should consider enabling that setting in the tracker to ensure as random ID generation as possible.