Help Understanding Remote Tag Providers & MQTT Engine

We’re in the process of implementing MQTT in our Ignition systems, but I’ve realized I might not fully understand how things work on the Engine side.

We currently have a single backend Ignition gateway with multiple sites connected. These sites are already in production, using Perspective HMIs populated by remote tag providers in our current setup.

We connected a test site to our new environment using MQTT, and I noticed that all the tags coming into the Engine are being created within the Engine tag provider. The problem is, our HMIs have hundreds of bindings that point to tags using the [site-name]/path/to/tag structure. I was under the impression that I could maintain this setup, with a tag provider subscribing to the relevant topics for each site.

Can someone help clarify how this should work and what steps I need to take?

We have other customers that use reference tags for this purpose. Would that work for you?

I’m not sure you can write back through reference tags. Also we are talking in some cases hundreds of thousands of tags.

You definitely can write to reference tags and those writes will result in writes to the MQTT Engine tags. Most people that do this script creation of the reference tags. Might that work for you? I should also note, if you are historizing data, there are some caveats to consider with reference tags. See this doc for more details: Configuring history on MQTT Engine tags - MQTT Modules for Ignition 8.x - Confluence