Description
Currently cbstats only gets the aggigated timing histograms for front end ops for our stats.log. This isn't super useful when analysing a cluster's stats.log when a cluster has multiple buckets, as it makes it hard to understand what front end ops are being received for a specific bucket and how that translates to the stats from cbstats which are collected on a per-bucket level.
To fix this, add functionality to mctimings to be able to display timings histograms on a per-bucket level. Then update cbcollect_info to use this functionality.