Using mod_wsgi-4.4.6.
Thanks for that information, that could be useful. I suppose there is no
external way to ask Apache how many workers are currently handling mod_wsgi
python requests or figure out the total mod_wsgi thread count busy with
requests?
Also, only partly off-topic: if running in daemon mode like this:
WSGIDaemonProcess rarch processes=30 threads=8 inactivity-timeout=1800
display-name=%{GROUP} python-eggs=/home/app/tg2env/lib/python-egg-cache
Is there any substantial advantage of worker MPM over prefork? It seems to
my understanding that daemon mode overcomes the limitations of prefork.
Thanks again!
On Thu, Dec 3, 2015 at 6:52 PM, Graham Dumpleton <[email protected]
> wrote:
>
> On 3 Dec 2015, at 11:48 PM, Kent Bower <[email protected]> wrote:
>
> Just trying to determine capacity utilisation for the process/server.
> Thank you!
>
> On Wed, Dec 2, 2015 at 4:21 PM, Graham Dumpleton <
> [email protected]> wrote:
>
>>
>> On 3 Dec 2015, at 2:03 AM, Kent <[email protected]> wrote:
>>
>> Is there an external way to detect whether a thread is currently handling
>> a request or how many threads are currently handling requests?
>>
>>
>> In the first instance have a look at:
>>
>>
>> https://code.google.com/p/modwsgi/wiki/DebuggingTechniques#Extracting_Python_Stack_Traces
>>
>> If don’t need so much detail and just after counts to determine capacity
>> utilisation for the process/server, then there are ways of getting such
>> metrics out of mod_wsgi itself.
>>
>> What I can’t remember right now is whether this specific metric is
>> available in latest released version or still sitting on a branch with a
>> bunch of other new metrics changes.
>>
>> What is the specific problem you are trying to solve?
>>
>
> What version of mod_wsgi are you using?
>
> After checking, the more recent versions (not the much older versions on
> most Linux distros) do have the metric counter I mention.
>
> The process metrics are available using:
>
> import mod_wsgi
>
> current_metrics = mod_wsgi.process_metrics()
>
> They include something like:
>
> mod_wsgi.process_metrics: {'cpu_system_time': 0.009999999776482582,
> 'request_busy_time': 0.00033, 'current_time': 1449186182.184397,
> 'memory_max_rss': 9646080L, 'memory_rss': 9646080L, 'pid': 5485,
> 'restart_time': 1449186179.973325, 'request_count': 0L, 'cpu_user_time':
> 0.03999999910593033, 'running_time': 2L}
>
> If you poll this at 1 second interval from a background thread, then
> capacity utilisation is calculated using:
>
> (current_metrics[‘request_busy_time’] -
> last_metrics[‘request_busy_time’]) / (current_metrics[‘current_time’] -
> last_metrics[‘current_time’])
>
> Personally I have been feeding this as well as a bunch of other stuff from
> unreleased version in a mod_wsgi branch into InfluxDB and charting it using
> Grafana. I should probably finally look at merging the branch so that
> people can play with it. Just not 100% sure I may want to tweak it a bit
> yet.
>
> Graham
>
> --
> You received this message because you are subscribed to a topic in the
> Google Groups "modwsgi" group.
> To unsubscribe from this topic, visit
> https://groups.google.com/d/topic/modwsgi/efdtI39ccRI/unsubscribe.
> To unsubscribe from this group and all its topics, send an email to
> [email protected].
> To post to this group, send email to [email protected].
> Visit this group at http://groups.google.com/group/modwsgi.
> For more options, visit https://groups.google.com/d/optout.
>
--
You received this message because you are subscribed to the Google Groups
"modwsgi" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to [email protected].
To post to this group, send email to [email protected].
Visit this group at http://groups.google.com/group/modwsgi.
For more options, visit https://groups.google.com/d/optout.
