Disk Allocation Threshold

Currently running into issues where nodes in cluster end up with disk imbalanced (10%+ plus difference between nodes with 3tb e.g. 300gb difference). This is with disk high and low watermark set to same.

Wondering there are any issues setting high watermark lower then low watermark. Shard indicies daily with UTC rollover so trying to avoid one node getting stuck all of the days active shards. That's why we'd prefer using high watermark which hopefully would only reallocate previous days shards to balance out cluster disk