Multiuser export via MQTT

[vladwing]

Currently openscale-sync sends all the data to the same topic in MQTT, making multi-user syncs impossible from the same phone/instance of openscale/openscale-sync. Would you be open to accepting a PR for this feature? Do you have any requirements for the design, in case you accept this feature?

[oliexdev]

Thanks for the suggestion — I went ahead and implemented multi-user support. It turned into a bit more work than
expected, because it grew into a larger sync rework rather than a small MQTT tweak.

For MQTT specifically:

• The userId is now part of the topic path (openScaleSync//measurements/), so each user gets their own
  topics and multi-user syncs from one phone work.
• The username is included in the JSON payload and used as the Home Assistant device/display name.
• Home Assistant discovery is generated per user (userId-suffixed unique_id/identifiers and per-user state_topic),
  following the usual HA/Zigbee2MQTT convention (stable id in the topic, name as display).
• The settings screen shows the resolved per-user topic path with a copy button.

Multi-user is a per-backend capability: MQTT, Webhook and InfluxDB sync all users, while Health Connect and Wger stay
single-user with a per-service user picker.

It would be great if you could test it. Two important notes:

1. Please use the latest dev build of openScale-sync.
2. You also need the matching dev build of openScale, not the Play Store/F-Droid release — the new sync app requires the
   updated openScale (there's a version gate that blocks syncing against an older openScale).

Looking forward to your feedback!

[vladwing]

Thank you so much!

I had one issue setting up testing for this. I installed the debug versions on the same phone where I run the standard ones, but openscale-sync-debug tries to connect to openscale-standard and complains about the version like you said. I will setup an emulator for testing, unless you have a better idea how to do it.

[vladwing]

I ended up backing up the standard version of the openscale, uninstalling all and the restoring the backup to the debug version. Now the sync works and I was able to confirm that new values are synced to HA through MQTT.

The bit that is unexpected to me is that only one user is updated in the new entities in HA - the one having the last measurement. This is not going to be a problem when using the app day to day, but for a use-case like this where you sync for the first time it's not great. Looking in the Mqtt log, I think the app sends the auto-discovery payloads for all users (confirmed because the entities are created in HA), but it sends only the last measurement. After pressing the "sync MQTT" button a toast saying "87 measurements synced" appears which made me think that all measurements were sent one by one, when in fact only the last one is sent.

[oliexdev]

This turned out to be a real multi-user bug, now fixed. Each openScale user now gets its own values published to its own retained …/last topic, so all users' entities show data in Home Assistant (not just the one with the latest measurement).

On top of that I added a retained history topic (openScaleSync/<userId>/measurements/history) that carries each user's full series, plus a Home Assistant blueprint to backfill it into long-term statistics — useful exactly for your first-sync use case. Everything is documented here:

👉 https://github.com/oliexdev/openScale-sync/wiki/MQTT-support

Could you grab the latest dev build and test again? Let me know if the per-user values (and the new history topic) work as expected on your setup.

[vladwing]

On it! I'll let you know how it works

[oliexdev]

@vladwing did you time to test it?