github.com/muhammedhassanm/blockchain@v0.0.0-20200120143007-697261defd4d/sawtooth-core-master/validator/sawtooth_validator/metrics/metrics.py

# Copyright 2018 Intel Corporation
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
#     http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
# ------------------------------------------------------------------------------
"""
Metric types
------------

Currently, three types of metrics are supported: gauge, counter, and timer.

- Gauge: Used to record a value that changes arbitrarily.
- Counter: Used to record a value that increments or decrements.
- Timer: Used to record the duration of tasks.

To add more metric types, corresponding mock metrics must be added to the end
of metrics.py as these mocks are used when metric reporting is disabled.

API Usage
---------

To create a new metric, a handle to the metrics collector must be created at
the beggining of the module with:

    from sawtooth_validator import metrics
    COLLECTOR = metrics.get_collector(__name__)

This creates a new handle which will tag all metrics created with the handle
using the full module name. To create a new metric, call the function with the
corresponding name:

    important_value_gauge = COLLECTOR.gauge("important_value")

The above creates a metric named
`"sawtooth_validator.module_name.important_value"` and tags it with the
hostname where the validator is running. If metrics reporting for this metric
is disabled, a mock object is returned which implements the same API as the
regular metric object. **Note:** Mock objects do not maintain state.

If the metric is part of a class, the instance should be passed in when the
metric is created like so:

    important_instance_value_counter = COLLECTOR.counter(
        "important_instance_value", instance=self)

This automatically adds the class name to the metric name.

Additionally, a metrics reporting level can be set and additional tags can be
added when it is created using the `level` and `tags` parameters:

    important_timer = COLLECTOR.timer(
        "important_timer",
        instance=self,
        level=metrics.DEBUG,
        tags={
            "name": self.name,
            "pid": os.getpid(),
            "instance": id(self),
        })

Tags should be used to separate metrics from multiple sources that are
collected in the same place. For example, `InstrumentedThreadPoolExecutor` uses
tags to distinguish threadpool metrics by the threadpool's name. While you
could also separate out these metrics by instance id, adding a name tag makes
interpreting the metrics much easier.
"""

import platform


DEBUG = 10
INFO = 20
PERF = 30
DEFAULT = INFO


def init_metrics(level=None, registry=None):
    """
    Initialize metrics reporting with the given level and registry. This
    should be called before get_collector().
    """
    MetricsCollector.set_instance(level, registry)


def get_collector(module_name):
    """
    Get a handle to the metrics collector.
    """
    return MetricsCollectorHandle(module_name.split(".")[-1])


class MetricsCollector:
    __instance = None

    @classmethod
    def set_instance(cls, level=None, registry=None):
        cls.__instance = cls(level, registry)

    @classmethod
    def get_instance(cls, level=None, registry=None):
        if cls.__instance is None:
            cls.set_instance(level, registry)
        return cls.__instance

    def __init__(self, level=None, registry=None):
        if level is None:
            level = DEFAULT
        self._level = level

        self._noop_registry = NoOpMetricsRegistry()
        self._registry = registry

        self._base_tags = (
            ("host", platform.node()),
        )

    def gauge(self, identifier, level, instance=None, tags=None):
        if self._registry is None or self._disabled(identifier, level):
            return self._noop_registry.gauge(identifier)

        return self._registry.gauge(
            self._join(identifier, instance, tags))

    def counter(self, identifier, level, instance=None, tags=None):
        if self._registry is None or self._disabled(identifier, level):
            return self._noop_registry.counter(identifier)

        return self._registry.counter(
            self._join(identifier, instance, tags))

    def timer(self, identifier, level, instance=None, tags=None):
        if self._registry is None or self._disabled(identifier, level):
            return self._noop_registry.timer(identifier)

        return self._registry.timer(
            self._join(identifier, instance, tags))

    # Private methods
    def _disabled(self, identifier, level):
        """Check if the metric is enabled based on the level."""
        return level < self._level

    def _join(self, identifier, instance=None, tags=None):
        """
        Join the identifier tuple with periods ".", combine the arbitrary tags
        with the base tags and the identifier tag, convert tags to "tag=value"
        format, and then join everything with ",".
        """
        tag_list = []
        if tags is not None:
            tag_list.extend(tags.items())
        tag_list.extend(self._base_tags)
        return ".".join(identifier) + "," + ",".join(
            "{}={}".format(k, v)
            for k, v in tag_list
        )


class MetricsCollectorHandle:
    def __init__(self, module_name):
        self._module_name = module_name

    def gauge(self, metric_name, level=DEFAULT, instance=None, tags=None):
        return MetricsCollector.get_instance().gauge(
            identifier=self._create_identifier(metric_name, instance),
            level=level,
            instance=instance,
            tags=tags)

    def counter(self, metric_name, level=DEFAULT, instance=None, tags=None):
        return MetricsCollector.get_instance().counter(
            identifier=self._create_identifier(metric_name, instance),
            level=level,
            instance=instance,
            tags=tags)

    def timer(self, metric_name, level=DEFAULT, instance=None, tags=None):
        return MetricsCollector.get_instance().timer(
            identifier=self._create_identifier(metric_name, instance),
            level=level,
            instance=instance,
            tags=tags)

    def _create_identifier(self, metric_name, instance=None):
        if instance is None:
            return (self._module_name, metric_name)
        return (self._module_name, instance.__class__.__name__, metric_name)


class NoOpMetricsRegistry:
    def __init__(self):
        self._noop_gauge = NoOpGauge()
        self._noop_counter = NoOpCounter()
        self._noop_timer = NoOpTimer()

    def gauge(self, identifier):
        return self._noop_gauge

    def counter(self, identifier):
        return self._noop_counter

    def timer(self, identifier):
        return self._noop_timer


class NoOpGauge:
    def set_value(self, *args, **kwargs):
        pass

    def get_value(self, *args, **kwargs):
        return 0


class NoOpCounter:
    def inc(self, *args, **kwargs):
        pass

    def dec(self, *args, **kwargs):
        pass


class NoOpTimer:
    def __init__(self):
        self._ctx = NoOpTimerContext()

    def time(self):
        return self._ctx


class NoOpTimerContext():
    def __enter__(self):
        pass

    def __exit__(self, exception_type, exception_value, traceback):
        pass

    def stop(self, *args, **kwargs):
        pass