On 03/18/2015 11:56 AM, Hank Beatty wrote:
Solved.
It appears the ticker_interval has to be set to 60 or maybe at least
increments of 60. Is that right?
No, that's not right. It's a good idea to pay attention to the relationship
between ticker_interval (which specifies how often the timer_event function
will be called) and the seconds per row in your circular buffer, but even if
those aren't correlated the filter shouldn't crash. If you're seeing that it
might be a bug.
There are a couple of global settings you might be bumping up against, though.
max_process_inject and max_timer_inject (see
http://hekad.readthedocs.org/en/v0.9.0/config/index.html#global-configuration-options)
limit the number of messages a single invocation of process_message and
timer_event can emit, and if you try to spit more than that out your filter
will be killed. I'm not sure if that's related, though, w/o seeing the rest of
your code.
I was trying a ticker_interval of 5 in an effort to get the graph faster.
In theory this should be possible, but if your seconds per row is > 5 seconds
then you're just going to see the current interval's number slowly grow, the graph
won't actually advance until the next row of data comes through.
-r
_______________________________________________
Heka mailing list
[email protected]
https://mail.mozilla.org/listinfo/heka
