We are looking at a project where we are going to have around 1100 sensors connection via 100 gateways to a TTN server. HomeAssistant will then utilize the built in TTN integration to pole for sensor data. The TTN server will be on the same network as the HA server (connected via Ethernet).
The question is will HA handle this? (we can physically tip whatever resources required into the HA server Disk/Memory/CPU Cores), is there any metrics available on this? We are measuring Temperature/Humidity and also alerting on these.
I think it’s the correct “channel” but I don’t know if anyone has any experience with this specific requirement to answer your question.
I would perform my own tests to see what limits I may encounter. For example: I would do a test setup to see how x hardware performs (cpu load, mem utilization, response time, etc.) with y number of sensors or gateways. Then scale up the sensors and/or hardware. Another approach would be to have multiple instances of hardware with Home Assistant to service a fixed number of sensors (ex. 5 HA servers to monitor 220 sensors or 20 gateways).
I agree with MaxK. The channel category if fine, but it’s unlikely that anyone has a comparable setup. Based on users who opt-in to sharing analytics, the TTN integration has 190 active users. Which isn’t all that surprising since TTN is mostly enterprise-focused, while HA is predominantly focused on home users.
LoRa-based devices, in general, are much less common among home users compared to WiFi, Zigbee, ZWave, and Bluetooth. TTNs subscription requirement may also be a significant detractor for many HA users who are looking for less cloud-dependency . While TTN is one of the better known LoRa-based networks, there are a number of other networks with differing degrees of focus on home automation that require no subscription.
