Sorry for the minsinfo guys. Been awhile since I looked at that code.
The delay parameter IS in ms.
The current tick resolution setting is in seconds, but I’ll chnage that in the next release.
Warnings still apply though about ticking too often and filling up your system queue with back events.
Is there any way of monitoring the queue size? Monitoring the queue size over time, and the composition queue (i e, N tasks of type A, M tasks of type B) seem like pretty useful profiling tools, for example. You could also write a system that balanced quality (degraded resolution) under load.
This falls into the general administration interface which we are working on in this phase.
And yes I agree queue statistics on a per slice as well as per user per slice are one of mant statistics that coudl be very useful!
administration interface - any news on this? cant seem to find it in version0.9r1966
Version 0.9 has some system monitoring logging that can be enabled with non-standard command line flags (the moral equivalent of -X, not guaranteed to stay the same between releases.) Its primitve and we intend to expand on it when we have time, but its what we’re using for stack and app tuning now.
I’ll ask the guys what they are and try to get a post up ASAP.
Jeff