I need some help figuring out why my home assistance instance (HASS OS VM) is using so much data.
I recently acquired a unifi UCG Ultra which gives me a pretty detailed breakdown of traffic flows in/out of my network. While looking at the dashboards I found that my HASS instance is the #1 traffic user on my network, by far.
Now like many of you here I do have a crippling Home Assistant addiction but I was not expecting that much traffic. How do I go about finding out what is causing this? My initial thought was CCTV cameras or something like that, it peaks every 45 minutes to 1 hour.
I don’t run anything “on top” that I can think of (for e.g. speedtests, frigate etc) that could be causing this (
The Glances Add-On shows transmit and receive MB/sec for each container (add-on), but it’s live data so you would need to watch it when you see those hourly peaks. My HAOS by comparison is using 4.39 Gig per month (upload and download combined) according to Unifi.
I think HASS is actually fine, but the USG is misreporting. My HASS Is a HAOS VM running on Proxmox, so I went to look at the traffic usage for the VM on proxmox itself.
The only thing I can think of that is generating this amount of data would be my IP TV cameras - I have several here on-site, and I also have some at a remote site that is connected by a wireguard tunnel but even if that were causing it I would expect to see a constant ~10mbps as its an RTSP stream that is on 24/7.
I assume your ucg reports for your entire LAN and the proxmox stats are for your ha vm only? In other words your spikes might be nothing to do with ha.
No - the stats from the UCG above are filtered by host (in this case my HASS VM only).
I also have a Observium setup (polling freq every 2 min) that is aimed at several devices around my network incl. my proxmox host where the HA OS VM runs. Below is what has gone in/out over the same period as the UCG Screenshots show week over its LAN interface. The spikes are to be expected, those are virtual machine backups. But it looks nothing like what the UCG reports.
UCG reports excessive data usage (to internet) from the HASS VM
Proxmox per-vm usage does not corroborate this
Observium does not corroborate this either
As I mentioned before the only thing I can think of that may be causing this is my frigate plugin. But then why would it be so peaky, and so regular? And why isn’t Observium showing it?
More investigation needed.
Will give glances a shot soon.
Anyone have any recommendations on a netflow collector that’s easy to set up? The UCG can feed the collector.
It would appear nothing is actually wrong. Earlier today I installed a netflow collector and sent flows from my UCG Ultra to it.
The UCG still reports that my home assistant is using 800M-1G/hour, but this is not corroborated by the netflow analyzer nor by the stats provided by the hypervisor (both these agree).
At this stage I’m going to just abandon this - seems like a bug in the Ubiquti firmware.
