Optimize redis writes for production

[odedfos]

Hi,
In my system I had a latency issue between my the web-server and the redis redis. This led me to discover some inefficiency in the metrics logic when updating metrics.

checkout
def metric(self, slug, num=1, category=None, expire=None)

in https://github.com/bradmontgomery/django-redis-metrics/blob/master/redis_metrics/models.py [331]

For each key a separate incrementation is performed which leads to multiple redis calls.
Better to use a pipe here and update all keys in a single api call.

for example:

...
pipe = self.r.pipeline()

pipe.incr(hour_key, num)
pipe.incr(day_key, num)
pipe.incr(week_key, num)
pipe.incr(month_key, num)
pipe.incr(year_key, num)

pipe.execute()

Cheers!

[bradmontgomery]

Hi @odedfos, this seems like a great idea. Do you mind making the change, and sending me a pull request, so I can merge? Be sure to add yourself to AUTHORS.rst as well.

Thanks!