GithubHelp home page GithubHelp logo

java-metrics's Introduction

Build Status Released Version

OpenTracing Metrics

The OpenTracing Metrics project enables any OpenTracing compliant Tracer to be decorated with support for reporting span based application metrics.

The project currently has support for reporting metrics via:

  • Micrometer
  • Prometheus (deprecated in favor of Micrometer)

A Tracer is decorated in the following way:

	Tracer tracer = ...;
	MetricsReporter reporter = ...;
	Tracer metricsTracer = io.opentracing.contrib.metrics.Metrics.decorate(tracer, reporter);

Metric Labels

Labels are used as a way to separate sampled metric values into related groups. A combination of label values will uniquely define a specific metric.

If one of the metric labels returned for a particular sampling point (i.e. span) returns a null value, then the metric will not be recorded. This provided a means to selective choose which metrics values are of interest.

For example, by default if the span.kind tag is not specified, it's label value will be null. This means that metrics for internal spans will by default not be recorded. If an application/service developer wants specific internal span metrics recorded, they can add a MetricLabel that returns an appropriate value for the span.kind for the spans of interest.

Label Types

Label types:

  • ConstMetricLabel

This label type returns a constant value, e.g. service name

  • OperationMetricLabel

This label type returns the span's operation name

  • TagMetricLabel

This label type attempts to obtain the value associated with the requested name from the span's tags, and if not found uses a default value.

  • BaggageMetricLabel

This label type attempts to obtain the value associated with the requested name from the span's baggage items, and if not found uses a default value.

Default Labels

By default, the metrics are reported using the following labels:

  • operation - the operation associated with the span

  • span.kind - the span.kind tag associated with the span, by default if not specified, then the metrics for the span will not be recorded

  • error - the error tag, by default the value will be false

Adding Metric Labels

An application may want to add specific labels to help classify the metrics being reported for each span.

For example, many Tracer implementations associate spans with a service name. This can also be achieved for the span metrics, by specifying a ConstMetricLabel when creating the reporter.

Customizing Metric Labels

When initializing the MetricsReporter, it would be possible to provide a MetricLabel for a default label, to override its value.

For example, a MetricLabel implementation could be provided for the error label, which could override the standard boolean value and potentially provide an alternative set of values based on other tags or baggage values associated with a span.

Reporting Metrics

Micrometer metrics reporting is provided by a specific implementation of the MetricsReporter interface.

For example,

// Your application needs to setup a concrete Micrometer backend
io.micrometer.core.instrument.Metrics.addRegistry(new SimpleMeterRegistry());

// Prepare a concrete OpenTracing tracer
Tracer tracer = getTracer();

// A reporter can then be created like this:
MicrometerMetricsReporter reporter = MicrometerMetricsReporter.newMetricsReporter()
    .withName("MyName")
    .withConstLabel("span.kind", Tags.SPAN_KIND_CLIENT)
    .build();

// Wrap the concrete Tracer, so that we can record the metrics about the reported spans
Tracer metricsTracer = io.opentracing.contrib.metrics.Metrics.decorate(tracer, reporter);

Builder methods are provided to enable new labels to be provided, or existing ones overridden.

Refer to the Micrometer documentation on how to get the metrics into a concrete backend, such as JMX, StatsD or Prometheus.

Reporting metrics with a Prometheus backend

Auto-configuration for Spring Boot applications of a Prometheus backend is provided via the module opentracing-metrics-prometheus-spring-autoconfigure. To auto-register an endpoint serving Prometheus metrics, export the property OPENTRACING_METRICS_EXPORTER_HTTP_PATH with the path to be used - e.g. "/metrics".

TracerObserver approach

Instead of decorating an OpenTracing tracer, it's also possible to combine the usage of Spring Boot's auto configuration feature and the TracerObserver from io.opentracing.contrib:opentracing-api-extensions-tracer.

Just include the artifact io.opentracing.contrib:opentracing-metrics-spring-autoconfigure into your Spring Boot application and the TracerObserver will be registered automatically.

Known Issues

Only works with ActiveSpanSource implementations that don't require a tracer specific Span implementation

The current mechanism uses a wrapper tracer implementation to identify when a span has finished. This requires a wrapped Span to be passed to the ActiveSpanSource.makeActive method, and therefore will fail if the ActiveSpanSource implementation has an expectation of receiving a particular Span implementation.

This wrapper approach is only being used as a short term workaround until a TracerObserver mechanism is available.

Release

Follow instructions in RELEASE

java-metrics's People

Contributors

jpkrohling avatar kevinearls avatar objectiser avatar yurishkuro avatar

Watchers

 avatar

Recommend Projects

  • React photo React

    A declarative, efficient, and flexible JavaScript library for building user interfaces.

  • Vue.js photo Vue.js

    ๐Ÿ–– Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.

  • Typescript photo Typescript

    TypeScript is a superset of JavaScript that compiles to clean JavaScript output.

  • TensorFlow photo TensorFlow

    An Open Source Machine Learning Framework for Everyone

  • Django photo Django

    The Web framework for perfectionists with deadlines.

  • D3 photo D3

    Bring data to life with SVG, Canvas and HTML. ๐Ÿ“Š๐Ÿ“ˆ๐ŸŽ‰

Recommend Topics

  • javascript

    JavaScript (JS) is a lightweight interpreted programming language with first-class functions.

  • web

    Some thing interesting about web. New door for the world.

  • server

    A server is a program made to process requests and deliver data to clients.

  • Machine learning

    Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.

  • Game

    Some thing interesting about game, make everyone happy.

Recommend Org

  • Facebook photo Facebook

    We are working to build community through open source technology. NB: members must have two-factor auth.

  • Microsoft photo Microsoft

    Open source projects and samples from Microsoft.

  • Google photo Google

    Google โค๏ธ Open Source for everyone.

  • D3 photo D3

    Data-Driven Documents codes.