FlowFuse
DevOps Pipelines: Support “Managed vs Local” flows on the same Node-RED instance (partial pipeline deployments)
Description
Summary In many ISA-95/OT scenarios, each Node-RED instance needs a mix of: 1. Standard, centrally managed flows (for publishing, normalization, visualization), and 2. Instance-specific “harvest” flows (machine/site adapters that vary per asset).
Today, DevOps pipelines assume the whole flow set is identical per environment/instance. We’d like pipelines to manage only a subset of flows while preserving/allowing local flows on the same instance.
Use Case Example
Consider 100 CNC machines: • Each one produces similar types of data (speed, status, production counts) • But each machine’s interface is slightly different and requires a custom “harvester” flow. • Once harvested, the data needs to pass through a common publishing and visualization pipeline — ideally managed and version-controlled through FlowFuse DevOps.
In the current model, deploying a new pipeline version risks overwriting local ingestion logic or forces separate Node-RED instances per asset — both of which complicate scaling and governance. As an OT engineer, I want to deploy a standard publishing/normalization package to all Node-RED instances via a FlowFuse pipeline, while keeping machine-specific ingestion flows local on each instance. When I promote a new version of the standard package, it updates only the managed flows—my local harvester logic remains untouched.
If there’s already a supported pattern that achieves partial management of flows (for example, deploying selected tabs or combining pipeline and local content), please point me in the right direction. Otherwise, this could be a strong enhancement for hybrid OT/IT use cases where Node-RED nodes act as data harvesting edges.
Which customers would this be available to
All customers. it’s a platform-level enhancement, not a gated enterprise feature.
