Hi Kireet, Have you had a look at Hannibal (https://github.com/sentric/hannibal)? It graphs the distribution of a table's regions across your cluster as well as region sizes and may point to unevenly distributed data which you could then try to correlate to load on a particular node.
-Eric On Wed, Jul 3, 2013 at 6:32 AM, Kireet <[email protected]> wrote: > Hello, > We are seeing some increased load on our system and we are trying to > determine why. We have historical read/write request count data (from jmx > metrics) but it's hard to pick out a definitive correlation between a > particular table and overall system load. Is there a way to tell which > table is causing the most i/o load on data nodes? For example we do a large > amount of 1k reads on one table but those counts have stayed relatively > flat. > > Thanks > Kireet > > -- *Eric Chang* Technology Lead, Data Services * Opower * <http://www.opower.com/> * ** We’re hiring! See jobs here <http://www.opower.com/careers> *
