Features

• automatic collection of request latency metrics for RPC methods, event handlers and HTTP requests

• custom application metrics are also exposed automatically

• HTTP endpoint exposing metrics to be scraped by Prometheus

Installation

pip install nameko-prometheus

You can also install the in-development version with:

pip install https://github.com/emplocity/nameko-prometheus/archive/master.zip

Usage

Define your application metrics using types from prometheus_client, such as Counter, Gauge, Histogram, etc. Add a PrometheusMetrics dependency provider to your service class. Then add a HTTP entrypoint (usually routed under /metrics URL) that forwards the request to expose_metrics() method of the provider. That’s it!

from nameko.rpc import rpc
from nameko.web.handlers import http
from nameko_prometheus import PrometheusMetrics
from prometheus_client import Counter


work_units = Counter(
    "my_service_work_units_total", "Total number of work units", ["work_type"]
)


class MyService:
   metrics = PrometheusMetrics()

   @rpc
   def say_hello(self):
      work_units.labels(work_type="hard").inc()
      return "Hello!"

   @http("GET", "/metrics")
   def serve_metrics(self, request):
      return self.metrics.expose_metrics(request)

There are a few automatically defined metrics as well. See the documentation of PrometheusMetrics for more.

See also the full stack example in example/ directory in this repo.

Documentation

https://nameko-prometheus.readthedocs.io/

Development

To run the all tests run:

tox

Note, to combine the coverage data from all the tox environments run:

set PYTEST_ADDOPTS=--cov-append
tox

PYTEST_ADDOPTS=--cov-append tox

Development roadmap

Things we’d like to have in the future:

• automatic registration of /metrics HTTP endpoint

• decorator to exclude specific methods from tracing

Authors

nameko-prometheus is developed and maintained by Emplocity.

License

This work is released under the Apache 2.0 license.