Molasses-server
includes the Node (with TypeScript support) SDK for Molasses. It allows you to evaluate a user's status for a feature. It also helps simplify logging events for A/B testing.
The SDK uses SSE to communicate with the Molasses Application. Once initialized, it takes microseconds to evaluate if a user is active. When you update a feature on Molasses, it sends that update to all of your clients through SSE and users would start experiencing that change instantly.
npm install @molassesapp/molasses-server
yarn add @molassesapp/molasses-server
To Require:
1import { MolassesClient } from "@molassesapp/molasses-server";2// or3const { MolassesClient } = require("@molassesapp/molasses-server");
Start by initializing the client with an APIKey
. This begins the polling for any feature updates. The updates happen every 15 seconds. This can be overriden with the refreshInterval
field. If you decide not to track analytics events (experiment started, experiment success) you can turn them off by setting the sendEvents
field to false
1// Initialize with your API Key2const client = new MolassesClient({3APIKey: "testapikey",4});56// Then init which is a promise7await client.init();
You can call isActive
with the key name and optionally a user's information. The ID field is used to determine whether a user is part of a percentage of users and uniquely identies a user. If you have other user segments based on user parameterss you can pass those in the params
field.
1// Once initialized you can start calling isActive2client.isActive("FOO_TEST", {3id: "123",4params: {5isBetaUser: "true",6},7});
You can check if a feature is active for a user who is anonymous by just calling isActive
with the key. You won't be able to do percentage roll outs or track that user's behavior.
1client.isActive("FOO_TEST");
To track whether an experiment has started you can call experimentStarted
. experimentStarted takes the feature's name, the molasses User and any additional parameters for the event.
1client.ExperimentSuccess(2"GOOGLE_SSO",3{4id: "baz",5params: {6teamId: "12356",7},8},9{10version: "v2.3.0",11}12);
To track whether an experiment was successful you can call ExperimentSuccess
. ExperimentSuccess takes the feature's name, the molasses User and any additional parameters for the event.
1client.ExperimentSuccess(2"GOOGLE_SSO",3{4id: "baz",5params: {6teamId: "12356",7},8},9{10version: "v2.3.0",11}12);
new MolassesClient(options)
#Creates a new Molasses client. It takes a set of options.
Field | Required | Type | Description |
---|---|---|---|
apiKey | required | string | The API Key for the current environment |
URL | optional | string | the base URL for Molasses |
autoSendEvents | optional | boolean | Whether to automatically send analytic events back to Molasses. Defaults to false . |
refreshInterval | optional | number | If Polling, how often Molasses should fetch the updated configuration. Defaults to 15000 , |
streaming | optional | boolean | Whether Molasses should use SSE or polling. Defaults to true |
init()
#Fetches configuration of features and begins polling
isActive(featureKey: string)
#Will check if feature is active for ALL users
isActive(featureKey: string, user: User)
#Will check if the feature is active for this particular based on the segment they are in.
experimentStarted(featureKey: string, additionalInformation: string[string], user: User)
#Will send an event started message when a user starts an a/b tests. This includes, whether user was in the experimental group (control or experiment), the experiment that was being tested and other metadata. If you want to include additional metadata use the additionalInformation
argument.
experimentSuccess(featureKey: string, additionalInformation: string[string], user: User)
#Will send an event success message when a user completes a set goal. This includes, whether user was in the experimental group (control or experiment), the experiment that was being tested and other metadata. If you want to include additional metadata use the additionalInformation
argument.
stop()
#Stops polling