Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Already on GitHub? Sign in to your account

add system load avg info to stats #216

wants to merge 1 commit into


None yet
2 participants

prymitive commented Apr 7, 2013

When looking at stats from multiple nodes it is useful to have also load avg stats from those nodes, so it would be nice to add it.

I'm not sure about availability, it seems that this is present on all UNIXes and Cygwin emulates it, but I've seen mixed reports so (provided that you are willing to merge it) we might:
a) push it as is and wait for people to complain that it does not compile on some platforms
b) ifdef it with __linux__, __FreeBSD__, __sun__ and __APPLE__ (for example)


unbit commented Apr 8, 2013

i have committed the first tech draft for the metric subsystem:


i think i will make a first version in 1.9.7 (if not in 1.9.6). I need only to define how to correctly manage time (the problem we had with carbon metrics with 1 second frequency)

So, i think i can integrate this directly as a metric


prymitive commented Apr 8, 2013

how to correctly manage time

AFAIR the issue was that we need metric fetching (reading metrics from source and storing it in uWSGI buffer) and pushing (sending to carbon/mongo/whatever) to not block each other, the most obvious solution would be 2 threads and fixed size queue in between them.


unbit commented Apr 8, 2013

It seems high resolution timers are not needed for the metrics thread, the problem rises when we send the values to a system that need to make computation on them (based on time). The metrics thread is only a collector. This should speedup development :)

@prymitive prymitive referenced this pull request Nov 14, 2013


System metrics #455


unbit commented Feb 9, 2014

Same as #236

@unbit unbit closed this Feb 9, 2014

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment