Owl 0.2.1

Monitor Falcon with Riemann

Owl

Measure how long your Falcon requests take and send each measurement as an event to Riemann.

Example

Owl is designed to be a mix-in class for API. To use Owl you create your own API class providing configuration parameters for Owl.

In Python 3 this could look something like this:

from falcon.api import API
from riemann_client.client import QueuedClient
from owl import Owl

class MonitoredAPI(Owl, API):

    def __init__(self, *args, **kwds):
        kwds["get_riemann_client"] = lambda: QueuedClient()
        kwds["owl_service"] = "MyAPI"  # service name for Riemann
        super().__init__(*args, **kwds)

api = MonitoredAPI()
# api.add_route(...)
# ...

In Riemann you get the end point and the request result status code as tags. To convert them into their own InfluxDB tags ( endpoint` and ``status-code) for later use (for example in Grafana) you could do something like this in your Riemann InfluxDB configuration:

(defn- parse-api-event
 [event]
 (if (contains? event :tags)
   (let [[endpoint status-code & other] (:tags event)]
     (assoc event
       :endpoint endpoint
       :status-code status-code))
   event))

(let [index (index)]
 (streams index
   ; Write API monitoring to database
   (where (service "MyAPI")
     #(info (parse-api-event %))
     (comp api-db parse-api-event))))

Check out the event_builder parameter if you want to customize the event parameters (change what service means or add a description for example).

Installation

$ pip install owl

Changelog

0.2.0

• generate a message with status 500 when a request throws an exception