Prometheus

Prometheus is a monitoring solution we use at Q2 for keeping track of the stats that matter most to a running service.

There are a few gotchas in working with Prometheus that we’ve attempted to abstract away with this module. To use it will be something like this:

prometheus.get_metric(
    MetricType.Counter,
    'caliper_endpoints',
    'Number of times called',
    labels={'endpoint': self.extension_name}
).inc()

enum q2_sdk.core.prometheus.MetricType(value)[source]

Available metric types according to Prometheus

Valid values are as follows:

Counter = <MetricType.Counter: <class 'prometheus_client.metrics.Counter'>>

Histogram = <MetricType.Histogram: <class 'prometheus_client.metrics.Histogram'>>

Info = <MetricType.Info: <class 'prometheus_client.metrics.Info'>>

Summary = <MetricType.Summary: <class 'prometheus_client.metrics.Summary'>>

Gauge = <MetricType.Gauge: <class 'prometheus_client.metrics.Gauge'>>

Enum = <MetricType.Enum: <class 'prometheus_client.metrics.Enum'>>

Metric = <MetricType.Metric: <class 'prometheus_client.metrics_core.Metric'>>

q2_sdk.core.prometheus.get_metric(metric_type, name, documentation, labels, namespace='', subsystem='', unit='', registry=<prometheus_client.registry.CollectorRegistry object>, chain=None, is_parent=False, **kwargs)[source]

Will return an instance of metric_type with given parameters as inputs, but will return an existing one if it already exists, creating a new one only if it’s the first time that metric combination has been seen.

Any key/value pairs in kwargs will be passed through to the underlying MetricType in prometheus_client.metrics (such as buckets for Histogram)

Return type:: MetricWrapperBase