This library pushes your metrics collected with statman to stathat.
Statman metrics maps to stathat data points in the following way:
- histogram -> value
- counter -> count
- gauge -> value
Statman keys are flattened and join with a "/", e.g {my, key} to "my/key".
Pretty much a carbon copy of newrelic-erlang.
Your stathat ez_key needs to be set as an application
environment variable for the hatman app.
You may also explicitly whitelist stats by setting the whitelist
application environment variable for the hatman app. If the whitelist
is undefined all metrics are sent to stathat.
So an example config would be something like this (whitelist is optional, if undefined it means all statman keys are considered whitelisted):
> application:get_env(hatman, ez_key).
{ok, "stathat@example.org"}.
> application:get_env(hatman, whitelist).
{ok, [{db, write_latency}, {db, read_latency}, <<"some_other_statman_key">>]}.
To get started and configuring which polling function to use, do this:
application:set_env(hatman, ez_key, "stathat@example.org").
application:start(hatman).
statman_aggregator:start_link().
hatman_poller:start_link(fun hatman_statman:poll/1).
To test that hatman works, do this:
application:start(hatman).
EzKey = "stathat@example.org".
hatman_stathat:ez_json(EzKey, hatman_stathat:sample_stats()).
I only need it for statman but if you get your metrics from some other tool it
should be fairly easy to plug in your on poll function which formats your data
for stathat. Check out how formating is done in hatman_statman.erl.