Quickstart for Agent

6 min

https://www.youtube.com/embed/kwNVdSXMGX8

Watch this 4-minute video on implementing Optimizely Agent with example microservices or follow the instructions to Configure Optimizely AgentConfigure Optimizely Agent on your preferred operating system.

Agent REST APIs

The Agent REST APIs are defined through OpenAPI (formerly known as Swagger) and can be found on localhost or in the developer documentationdeveloper documentation.

Note The following examples use Python code for brevity, but can be expanded for any programming language.

Start an HTTP session

Each request made to Optimizely Agent is in the context of an Optimizely SDK key. SDK keys map API requests to a specific Optimizely Feature Experimentation projectproject and environmentenvironment. You can configure a global request header by using the requests.Session object.

import requests
s = requests.Session()
s.headers.update({'X-Optimizely-SDK-Key': '<YOUR_SDK_KEY>'})

To authenticate, pass your SDK key as a header named X-Optimizely-SDK-Key in your API calls to Optimizely Agent. You can find your SDK key in app.optimizely.com under Settings > Environments > SDK Key. Remember you have a different SDK key for each environment.

ImportantYou must start an HTTP session. Future examples will assume this session is being maintained.

Get the current environment configuration

The /config/config endpoint returns a manifest of the current working environment.

resp = s.get('http://localhost:8080/v1/config')
env = resp.json()

for key in env['featuresMap']:
    print(key)

Run a feature flag rule

The /decide?keys=/decide?keys= endpoint decides whether to enable a feature flag or flags for a given user. You can decide on multiple flags with this syntax. /v1/decide?keys=flagA&keys=flagB. Optimizely provides a userId through the request body. The API evaluates the userId to determine which flag rule and flag variation the user buckets into.

Rule types include the following:

A/B testsA/B tests – Flag variations are measured against one another.

Targeted deliveryTargeted delivery – Progressively make the flag available to the selected audienceaudience.

This endpoint returns an array of `OptimizelyDecision objects`OptimizelyDecision objects, which contains information about the flag and rule the user was bucketedbucketed into.

params = { "keys": "flagA" }
payload = {
    "userId": "test-user",
    "decideOptions": [
        "ENABLED_FLAGS_ONLY",
        "INCLUDE_REASONS"
    ],
    "userAttributes": {
        "attr1": "sample-attribute-1",
        "attr2": "sample-attribute-2"
    }
}
resp = s.post(url = 'http://localhost:8080/v1/decide', params=params, json=payload)

print(resp.json())

The decide APIdecide API is a POST to signal to the caller that there are side-effects. This endpoint results in a decision eventdecision event sent to Optimizely to analyze the A/B test results on the Optimizely Experiment Results pageOptimizely Experiment Results page. By default, a decision is not sent if the feature flag is part of a targeted delivery.