falcon-telegraf-middleware

What is this?

Package to ease gathering metrics from Falcon APIs into InfluxDB via a Telegraf. By use of middlewares this orthogonal functionality is added in a hassle-free way minimizing required work. Sane defaults make quickstart really quick 👍

How?

Inject the middleware (or couple of them) when creating Api instance:

from falcon_telegraf import LogHits, Timer

telegraf_client = TelegrafClient('localhost', 8094)
app = falcon.API(
    middleware=[
        LogHits(telegraf_client),
        Timer(),
    ]
)

Note that passing telegraf_client is totally optional. Measurement names can be set when creating middleware, prefixed or autogenerated.

Middlewares

LogHits

Creates a metric named hits-path/to/endpoint with one fields called hits with value 1. Executes after on-* method was called. Because of this you can add custom tags and values. Example:

def on_get(self, request, response, some_path_variable):
    request.context['telegraf_tags']['my_new_tag'] = 'this_tags_value'
    request.context['telegraf_values']['some_path_variable'] = some_path_variable
    ...

Timer

Reports time in ms elapsed between registering response and registering request. Default metric prefix is time-. Other things in processing pipeline can affect it's readings so take them with a grain of salt.

Testing

pip install -e .
pytest

Deploy

First bump version in setup.py in master branch via merge request. Afterwards push a version tag and wait.

git tag `date +"%Y.%-m.%-d.1"`
git push --tags