[erlang-questions] start_timer and timer fired lag
Tue Mar 27 14:57:09 CEST 2018
On 2018-03-27 08:36 +0000, Alexander Petrovsky wrote:
> Leo, I think Jesper talk about
On 2018-03-27 12:29 +0000, Jesper Louis Andersen wrote:
> Yes, that is the thing I'm talking about. Apart from long_schedule, I'd
> also enable the other kinds of monitors, especially long_gc. Usually, when
> the system reports one of these things, it is worth investigating why it is
> doing so. A timer might also arrive late because a poor process has way too
> much GC work to do, while also having too few cores. So having a marker for
> this tend to be good infrastructure design.
> IIRC, Fred's book "Erlang in Anger" also touches on this (but haven't
> looked it up, I must admit).
Thanks for the pointers. yes, EiA did touch on it.
More information about the erlang-questions