On 10/6/05, Peter Bodik <[EMAIL PROTECTED]> wrote: > Hi, > we're a research group at UC Berkeley starting a new project and we're > thinking of using ganglia for part of it. We use Rubis -- an open-source > eBay-like J2EE application -- running on 1-5 servers (webserver, JBoss, > Rubis + a database) and we need to monitor the whole system. We want to > monitor the standard metrics (CPU, mem, IO, ...), but we're mainly > interested in the applications. We'll heavily instrument the applications to > report a lot of metrics. For example, each Java Bean would report some > statistics about its usage, how long it takes to execute requests, number of > warnings, errors, exceptions and so on. We'd like to get all the data to one > machine and do some analysis in real-time. For some of the metrics we'll > need to get updates every second. > > Does anybody have experience with such setup? Basically, the number of hosts > would be pretty small, but the number of metrics could be large (100s or > 1000s) and for some of them we'll need very fast updates.
While it would be nice to have per-process or per-application metrics, I believe it was recently stated in an e-mail just sent in the last 24 hours about a similar task that it's not currently available in Ganglia, although it might be a nice feature to put into the libmetrics codes. I could be wrong though, and someone else may have a better idea of the situation. -- ~Mike - Just my two cents - No man is an island, and no man is unable.
