metricslib 0.3.0

Metrics collection library

Metricslib is a package that can be used to send application metrics to metric collections services.

Supported metric collection services

• statsd

Installation

Metricslib requires python >= 3.5. Install the latest version using pip

pip install metricslib 

Usage

Metricslib provides a decorator that can be used on a function where we want to collect metrics about how many times it was called, how many times it executed successfully, how many times it was executed with errors and how long it took to run.

For the moment the only supported metric collection service is Statsd.

from metricslib.utils import configure_metrics_from_dict
from metricslib.decorators import capture_metrics

@capture_metrics(
    request_metric="myapp.do_something.request",
    error_metric="myapp.do_something.error",
    success_metric="myapp.do_something.success",
    execution_time_metric="myapp.do_something.execution"
)
def do_something():
    print("hello world")


@capture_metrics(
    request_metric="myapp.do_something.request",
    error_metric="myapp.do_something.error",
    success_metric="myapp.do_something.success",
    execution_time_metric="myapp.do_something.execution"
)
def do_something_bad():
    raise Exception()


def main():
    config = {
        "STATSD_HOST": "localhost",
        "STATSD_PORT": 8125
    }
    configure_metrics_from_dict(config)

    do_something()

    # we want this function to raise an exception in order to test the error
    # metric
    do_something_bad()


if __name__ == "__main__":
    main()

Instead of using the decorator you can create counter objects.

from metricslib.config import configure_metrics_from_dict
from metricslib.utils import get_metrics


def main():
    config = {
        "STATSD_HOST": "localhost",
        "STATSD_PORT": 8125
    }

    configure_metrics_from_dict(config)

    metrics = get_metrics()
    counter = metrics.counter("myapp.count")
    counter.incr()


if __name__ == "__main__":
    main()

You can also measure the time duration of an operation.

from time import sleep

from metricslib.config import configure_metrics_from_dict
from metricslib.utils import get_metrics


def main():
    config = {
        "STATSD_HOST": "localhost",
        "STATSD_PORT": 8125
    }

    configure_metrics_from_dict(config)

    metrics = get_metrics()
    duration = metrics.duration("myapp.time")

    duration_measurement = duration.begin()
    sleep(2.0)
    duration_measurement.end()


if __name__ == "__main__":
    main()