cbclusterstats - add a new cli which runs "stats all,tap,checkpoint" stats against all nodes in the cluster and captures certain information about cluster's state

the high level information

how many nodes
how many buckets
how much ram per bucket
mem_used per bucket
ops/sec on the bucket per node
gets/sets ops per/sec
important timings stats from memcached ( disk fetch , disk commit , bg_wait)
miss ratio
dgm ratio
resident ratio ( active versus replica)

again we dont want to spit out 100 stats here and we want to be selective and only display those stats are quite important for troubleshooting

if we want more stats we can always look at the diags and get the full snapshot over time