Thanks Matt. - There are 80 components on canvas. - NiFi is not clustered. - DEBUG level for org.apache.nifi.web.filter.TimerFilter shows request duration somewhere between 2 and 20 millis.
-- Jagrut On Mon, Oct 10, 2016 at 12:26 PM, Matt Gilman <[email protected]> wrote: > Jagrut, > > There are a number of things we can look at in order to track down the > ultimate cause. A couple questions > > - How many components are on the canvas? As the number of components > rendered on the canvas increase, the more resources your browser is going > to be using. > - Is your NiFi instance clustered? If so, the request must go to the NCM > and then replicated throughout the cluster before returning the response to > the browser. While these requests are replicated concurrently, this means > that each request is only going to be as fast as your slowest node (in > terms of handling user requests). > > There is some logging which we can enable to track the response times for > each request. If you set debug level logging for org.apache.nifi.web. > filter.TimerFilter in your conf/logback.xml you should see some messages > in your log detailing the request duration. > > With these details we should be able to help track down where the latency > is happening. > > Matt > > On Mon, Oct 10, 2016 at 3:15 PM, Jagrut Sharma <[email protected]> > wrote: > >> I am observing that moving a NiFi processor group on the canvas is taking >> a few seconds to complete. During this time, a spinning circle icon is >> visible beside the 'Stats last refreshed' time. >> >> NiFi version is 0.6.1. No flows are currently running, active threads = >> 0, queued = 0. >> >> Any pointers to debug the issue will be helpful. >> >> Thanks! >> -- >> Jagrut >> >>
