This repository contains a number of custom activity examples for Salesforce Journey Builder. These examples are intended to help you understand how Journey Builder works and get a head start on building your own custom activities.
Journey Builder is a marketing planning tool that integrates the various channels and services in the Salesforce ecosystem. The application empowers marketers to manage the customer life cycle by composing a Journey workflow on a drag-and-drop canvas.
Learn more about Journey Builder
The quickest way to get started is to install Node.js then run the app locally:
# Install package dependencies
npm install
# Run the Express app in development mode
npm run dev
A webapp will be available at http://localhost:8080/
You will be greeted with a landing page with links to the various examples we've created.
You can deploy this example to Heroku and start working with Journey Builder Custom Activities Today!
Example of a custom activity that utilizes an external service to generate a discount code where the user inputs the discount percent in the configuration.
Example of a Rest Decision Split where your application tells the contact which way to go through the journey.
We support the following execution parameters that allow you to configure the timeout and retry values Journey Builder should use when sending a request to the external web service that the custom activity will invoke.
It is possible to configure these values for each instance of a custom activity. Use config.js to show the config in the UI for configuring the custom activity and save it when saving the journey.
For details, please go to [Custom Activity Configuration(https://developer.salesforce.com/docs/atlas.en-us.noversion.mc-app-development.meta/mc-app-development/custom-activity-config.htm)]
execute.timeout - How long, in milliseconds, before each rest activity in the journey times out. Must be from 1,000 to 100,000 milliseconds. Default is 60,000 milliseconds.
execute.retryCount - How many times to retry each rest activity in the journey after the rest activity times out. Must be from 0 to 5. Default is 0.
execute.retryDelay - How long, in milliseconds, to wait before each rest activity in the journey is retried. Must be from 0 to 10,000 milliseconds. Default is 1,000 milliseconds.
execute.concurrentRequests - How many rest activities to run in parallel. Must be from 1 to 50. Default is 1, which means no concurrent requests. Before you use concurrent requests, test the scalability and performance of the target site. If you observe increased gateway errors or timeouts, consider adding retry and increasing the timeout value.
When retry count is bigger than 1, our system will retry the request in case of network error, gateway error (5XX) or timeout. It is recommended that the customer set timeout value to be a large value (such as 10,000 or 30,000) to avoid canceling requests by Journey Builder when server is still sending data. In case of retry, it might be possible that the external, to Journey Builder, service already processed the request, but a duplicate request was sent from Journey Builder due to the above errors and retry.
We recommend the developer of the web service the custom activity calls to implement dedupe logic. For each request, we include two Guid fields activityId and definitionInstanceId in the request body. Parse the values from the payload and use the combination of these two values as a way to dedupe a request. When Journey Builder is re-trying the request, the above two fields are the same, but for a different request, the values will be different.
The performance of a journey that has a custom activity is directly related to the round trip latency from Journey Builder's system to the external web service and the scalability of the external web service.
The Journey Builder team has observed custom activity implementations, which include the external web service the custom activity is calling, which can process as little as 20,000 contacts per hour. There are also performant implementations of custom activity and the associated web service they are calling which can process as much as 5 million contacts per hour.
To attain higher performance you will have to engage with your Salesforce account team such that you can get help tuning the paramaters that play a role in performance.
We would like to hear from you if you have questions about these examples or if you have ideas for other examples that you would like to see included. Just log a new issue and we'll our best to help!
Source code is licensed under BSD 3-Clause
