I am also curious what the results of this survey will be. At Berkeley. we use Chukwa to monitor varying-sized dev/test clusters on EC2, from perhaps 10 to 100 nodes. Mostly we need the real-time path to watch load, and the stored data and logs to do subsequent analysis. We use some MR jobs for analysis, and some standalone non-MR processes that read from HDFS.
--Ari On Tue, Feb 23, 2010 at 2:34 PM, Jerome Boulon <[email protected]> wrote: > Hi, > I would like to get some stats on how chukwa is used today? > - cluster status (personal,business(experiment,dev,test,prod) > - cluster size > - chukwa is used to do what? > -- UI > -- log collection > -- data analytics > ---- hive > ---- pig > ---- M/R > > Thanks in advance, > /Jerome. > > -- Ari Rabkin [email protected] UC Berkeley Computer Science Department
