0.4.0 • Published 2 months ago

@openfeature/open-telemetry-hooks v0.4.0

Weekly downloads
-
License
Apache-2.0
Repository
github
Last release
2 months ago

OpenTelemetry Hooks

The OpenTelemetry hooks for OpenFeature provide a spec compliant way to automatically add feature flag evaluation information to traces and metrics. Since feature flags are dynamic and affect runtime behavior, it’s important to collect relevant feature flag telemetry signals. These can be used to determine the impact a feature has on application behavior, enabling enhanced observability use cases, such as A/B testing or progressive feature releases.

Installation

$ npm install @openfeature/open-telemetry-hooks

Peer dependencies

Confirm that the following peer dependencies are installed.

$ npm install @openfeature/server-sdk @opentelemetry/api

Hooks

TracingHook

This hook adds a span event for each feature flag evaluation.

MetricsHook

This hook performs metric collection by tapping into various hook stages. Below are the metrics are extracted by this hook:

  • feature_flag.evaluation_requests_total
  • feature_flag.evaluation_success_total
  • feature_flag.evaluation_error_total
  • feature_flag.evaluation_active_count

Usage

OpenFeature provides various ways to register hooks. The location that a hook is registered affects when the hook is run. It's recommended to register both the TracingHook and MetricsHook globally in most situations, but it's possible to only enable the hook on specific clients. You should never register these hooks both globally and on a client.

More information on hooks can be found in the OpenFeature documentation.

Register Globally

The TracingHook and MetricsHook can both be set on the OpenFeature singleton. This will ensure that every flag evaluation will always generate the applicable telemetry signals.

import { OpenFeature } from '@openfeature/server-sdk';
import { TracingHook } from '@openfeature/open-telemetry-hooks';

OpenFeature.addHooks(new TracingHook());

Register Per Client

The TracingHook and MetricsHook can both be set on an individual client. This should only be done if it wasn't set globally and other clients shouldn't use this hook. Setting the hook on the client will ensure that every flag evaluation performed by this client will always generate the applicable telemetry signals.

import { OpenFeature } from '@openfeature/server-sdk';
import { MetricsHook } from '@openfeature/open-telemetry-hooks';

const client = OpenFeature.getClient('my-app');
client.addHooks(new MetricsHook());

Custom Attributes

Custom attributes can be extracted from flag metadata by supplying a attributeMapper in the MetricsHookOptions or TracingHookOptions.

In the case of the MetricsHook, these will be added to the feature_flag.evaluation_success_total metric. The TracingHook adds them as span event attributes.

// configure an attributeMapper function for a custom property
const attributeMapper: AttributeMapper = (flagMetadata) => { 
    return {
        myCustomAttribute: flagMetadata.someFlagMetadataField,
    };
};
const metricsHook = new MetricsHook({ attributeMapper });
const tracingHook = new TracingHook({ attributeMapper });

Development

Building

Run nx package hooks-open-telemetry to build the library.

Running unit tests

Run nx test hooks-open-telemetry to execute the unit tests via Jest.

0.4.0

2 months ago

0.3.0

7 months ago

0.2.4

9 months ago

0.2.3

10 months ago

0.2.2

10 months ago

0.2.1

10 months ago

0.2.0

10 months ago