how to instrument a bottle app about client_python

[sebgoa]

I managed to do it with tornado, but is there a best practice without tornado:

from tornado.wsgi import WSGIContainer
from tornado.ioloop import IOLoop
from tornado.web import FallbackHandler, RequestHandler, Application
from prometheus_client import start_http_server, Summary

from flask import Flask
from prometheus_client import Counter

app = Flask(__name__)

REQUEST_TIME = Summary('request_processing_seconds','Time Spent processing request')

@app.route('/')
@REQUEST_TIME.time()
def handler():
    return "hello world"

class MainHandler(RequestHandler):
    def get(self):
        self.write("This message comes from Tornado")

tr = WSGIContainer(app)

application = Application([
                          (r"/tornado", MainHandler),
                          (r".*", FallbackHandler, dict(fallback=tr)),
                          ])

if __name__ == "__main__":
    application.listen(5000)
    start_http_server(9999)
    IOLoop.instance().start()

[drawks]

Here is a working example of instrumenting a few routes in a basic bottle application and making the metrics route available an attached to the application. This also uses the MultiProcessCollector since it would be fairly atypical (at all but the smallest scale) to run a wsgi application as a single process. This requires that the prometheus_multiproc_dir variable be set. This can typically be accomplished in your wsgi runner, for example in uwsgi you could add the following to your config

env = prometheus_multiproc_dir=/tmp/helloworld.stats

from bottle import route, run, template, Bottle
from prometheus_client import multiprocess
from prometheus_client import generate_latest, CollectorRegistry, Gauge, Counter

application = Bottle()
registry = CollectorRegistry()
multiprocess.MultiProcessCollector(registry)

IN_PROGRESS = Gauge("inprogress_requests", "help", multiprocess_mode='livesum')
REQUESTS = Counter('http_requests_total', 'Description of counter', ['method', 'endpoint'])


@IN_PROGRESS.track_inprogress()
@application.route('/hello/<name>')
def index(name):
    REQUESTS.labels(method='GET', endpoint="hello").inc()
    return template('<b>Hello {{name}}</b>!', name=name)

@IN_PROGRESS.track_inprogress()
@application.route('/metrics')
def metrics():
    data = generate_latest(registry)
    return data

[drawks]

For sake of completeness, here is the equivalent example with none of the multiprocessing stuff:

from bottle import route, run, template, Bottle
from prometheus_client import multiprocess
from prometheus_client import generate_latest, REGISTRY, Gauge, Counter

application = Bottle()

IN_PROGRESS = Gauge("inprogress_requests", "help")
REQUESTS = Counter('http_requests_total', 'Description of counter', ['method', 'endpoint'])


@IN_PROGRESS.track_inprogress()
@application.route('/hello/<name>')
def index(name):
    REQUESTS.labels(method='GET', endpoint="hello").inc()
    return template('<b>Hello {{name}}</b>!', name=name)

@IN_PROGRESS.track_inprogress()
@application.route('/metrics')
def metrics():
    return generate_latest(REGISTRY)


application.run()

[sebgoa]

thanks a lot for your answer, I will try it asap.

[sebgoa]

Ok this is working fine. I just need to figure out how to export the metrics into my prometheus server which is running on a different node...