Hi,

  I would like to monitor in a given time interval (e.g., 1min, 5min,
15min) the utilization of the worker pool. I think that this can be
reflected by some moving average of how many tasks are either waiting or
being processed (like is done for CPU load in Unix systems). If this
average approaches the number of processes I know that I need to do
something.

  Alternatively, I could use a measure of what percentage of the time
interval there was an available worker.

  Note that CPU utilization of the machine is not enough, because some of
my tasks have a fair bit of IO.

  I suspect that this can be done with the stats, snmp, or metrics
subsystems, but the documentation is very sparse. I see things like "By
default uWSGI creates a lot of metrics (and more are planned)" (metrics
page). But where are these metrics explained.

Thanks,
George.
_______________________________________________
uWSGI mailing list
[email protected]
http://lists.unbit.it/cgi-bin/mailman/listinfo/uwsgi

Reply via email to