Thanks Tom. I was thinking that it's worth converting to ms simple because of the counters we have. Total processing time is a long, which has a max value of about 4 seconds when converted to ms. 4 seconds is very small and I'd imagine that as far as messaging total processing time on a particular function, we'll exceed that value quickly. In ms, we can get about 50 days worth of constant processing time until it rolls over.
Interesting article. I haven't seen that one before. On Sun, May 5, 2013 at 10:19 PM, Tom Cunningham <[email protected]> wrote: > > I think I read that System.nanoTime is a better way of measuring elapsed > time (regoogled it - > https://blogs.oracle.com/dholmes/entry/inside_the_hotspot_vm_clocks). My > thought was that we should store the more precise time, but if you're > looking to display it in a UI, I think it probably makes sense to convert to > milliseconds for display. > > > On 5/5/13 11:55 AM, Alex O'Ree wrote: >> >> It looks like its in the unit of nanoseconds. Is there a particular >> reason this was used over milliseconds? I'm having difficult imagining >> a use case where such granularity was required >> >> On Sat, May 4, 2013 at 3:57 PM, Kurt T Stam <[email protected]> wrote: >>> >>> On 5/3/13 10:12 PM, Alex O'Ree wrote: >>>> >>>> Does anyone know the unit of measure for the jUDDI JMX metrics? >>>> >>>> After accessing a few endpoints after firing up tomcat, I'm seeing >>>> numbers in the 750 million, which doesn't make a whole lot of sense >>>> unless there's some units associated with it >>> >>> It just increments with each invocation. It is supposed to start at 0 I >>> would think. Maybe it only gets initialized correctly on jboss. Tom >>> Cunningham should know. >>> >>> --K > >
