Reactive Programming with Salesforce Data

Summary

Salesforce Platform Events and Change Data Capture enable custom applications to react to activity in Salesforce, motivating an event-sourced design for responsive, durable integrations. The article demonstrates a Node.js app on Heroku that consumes the Salesforce Streaming API through jsforce 1.9.0 or later, whose Faye-based streaming client handles authentication, network connectivity, and message callbacks. It notes that a basic subscription loses its position on restart and cannot coordinate parallel processing, then uses Redis to persist each topic’s Replay ID and distribute messages through pub/sub or reliable queues. Because the Salesforce Streaming API lacks consumer-group delivery, a singleton stream_consumer feeds Redis while scalable web or worker processes handle work. The design also recommends a separate non-admin integration user and monitoring connection status, active clients, last message time, and processing duration.

Takeaways

1. 1

   jsforce 1.9.0 and later provides Salesforce Streaming API support through a Faye-based client that manages authentication and network connectivity.

2. 2

   Redis makes the stream consumer durable by storing the last consumed Replay ID and supports distribution to parallel processes through pub/sub or reliable queues.

3. 3

   The recommended deployment uses one Salesforce stream_consumer process, separate scalable web or worker processes, a non-admin integration user, and stream-health monitoring.