metrik

Metrics aggregator for logfmt-style metrics

Assumes a single source

Follows: [https://github.com/ryandotsmith/l2met/wiki/Usage](l2met guidelines)

count#

Sums count# style metrics and includes other key=value pairs.

Enable with -c

Emitted very -i seconds.

measure#

Logs count, sum, min, max, mean, variance, and stdev for a rolling window.

Enable with -m

Set tumbling window size with -w

Usage

Try it out:

> npm install -g metrik
> cat logs.txt
2013-10-25T20:19:41.739142+00:00 app[web.3]: count#vault_usage.http_201=1
2013-10-25T20:19:41.739403+00:00 app[web.3]: count#vault_usage.http_2xx=1
2013-10-25T20:19:41.739616+00:00 app[web.3]: measure#products_usage_events_close=0.070309922ms
2013-10-25T20:19:41.799430+00:00 app[web.14]: count#vault_usage.http_201=1
2013-10-25T20:19:41.799430+00:00 app[web.14]: count#vault_usage.http_2xx=1
2013-10-25T20:19:41.799577+00:00 app[web.14]: measure#products_usage_events_open=0.010993464ms
2013-10-25T20:19:41.826215+00:00 app[web.40]: count#vault_usage.http_201=1

count#

> cat logs.txt | metrik -c

count#vault_usage.http_201=3
count#vault_usage.http_2xx=2
units=ms sample#products_usage_events_close.mean=0.070309922 sample#products_usage_events_close.median=0.070309922 sample#products_usage_events_close.perc95= sample#products_usage_events_close.perc99= now=1383021881431 n=1
units=ms sample#products_usage_events_open.mean=0.010993464 sample#products_usage_events_open.median=0.010993464 sample#products_usage_events_open.perc95= sample#products_usage_events_open.perc99= now=1383021881431 n=1

measure#

Use it on your dyno:

web: bundle exec ruby app.rb | metrik -c worker: bin/worker | metrik -m