pyometer

Python 3 metrics inspired by Dropwizard Metrics

Features

Metric Keys

Metric keys are used to uniquely identify a single metric. They consist of a name (ordered tuple of values) and/or tags (dictionary of name/value pairs).

from pyometer import metric_key

# Metrics by name
key = metric_key(name=("project", "metrics"))
# Metrics by tag
key = metric_key(tags={"env": "production"})
# Metrics by name and tag
key = metric_key(name=("project", "metrics"), tags={"env": "production"})

MetricRegistry

The MetricRegistry is the central container of metrics for an application. It can be used to create or register metrics that will later be available to metric reporters.

from pyometer import MetricRegistry, metric_key

registry = MetricRegistry(
    # Metrics created with this registry will be prefixed with this name
    base_key=metric_key(name=("project", "metrics"))
)

ValueGauge

from pyometer import metric_key

# Manually set the value of a gauge with ValueGauge
gauge = registry.value_gauge(key=metric_key(name="queue_size"), initial_value=0)
gauge.set_value(100)

SupplierGauge

from pyometer import metric_key


# Automatically pull the value of a gauge with SupplierGauge
def get_queue_size():
    return 4


gauge = registry.supplier_gauge(key=metric_key(name="queue_size"), supplier=get_queue_size)

Timer

from pyometer import metric_key
from pyometer.decorator import timer


@timer(registry=registry, key=metric_key(name="get_users"))
def get_users():
    """ fetch users """

Counter

from pyometer import metric_key

counter = registry.counter(key=metric_key(name="created_users"))
counter.increment(4)
counter.decrement(2)

Reporters

GrafanaCloudGraphiteReporter

from pyometer.reporter import GrafanaCloudGraphiteReporter

# Create a reporter for a Grafana-hosted Graphite instance
reporter = GrafanaCloudGraphiteReporter(registry=registry,
                                        metrics_url="<metrics url>",
                                        instance_id="<instance id>",
                                        api_key="<api key>")

# Manually push metrics
reporter.report()

brollyed/pyometer