UPnP/IGD integration missed some entities after 2021.10 update

After updated to 2021.10 I missed 5 entities reported by my Tenda Nova mesh routers.
I’m currently on 2021.10.2 restarted severals times and deleted the integration and set up it again.

I missed all entities reporting */s data, what was specially useful to have an estimation of bandwidth consumed at home.

Entities reported before update

entities reported after update

I will appreciate your comments

Best, Facundo

Fwiw, I gave the same issue with arris modern.

P.s. I have three tendas in bridge mode, I couldn’t use the built in router since it doesn’t have static assigns for IP address

Hi, thanks for your prompt reply.
The 13 entities on first pic was working well until I updated to 2021.10. Nothing changed on my Tenda mesh network. Indeed I created new entities to translate KiB/s to Mbps as here below.

  • platform: template
    sensors:
    nova_mbps_down:
    friendly_name: “Nova Mbps Down”
    value_template: “{{ (states(‘sensor.nova_kib_s_received’) | float / 122.0703125) | round(1) }}”

    nova_mbps_up:
    friendly_name: “Nova Mbps Up”
    value_template: “{{ (states(‘sensor.nova_kib_s_sent’) | float / 122.0703125) | round(1) }}”

But as all */s envites are not reported anymore I have no more Mbps Up and Down indication in HA.

Is the same issue as reported in UPnP/IGD integration of UniFi Dream Machine

The internet network monitoring sensors are returned with very strange values. I’ve tried to convert these values to kbit/s with a converter many ways but the results never passed to the real values.

Yup, something is messed up. It hasn’t been right for the past two updates (even after applying the latest update).

My receive rate is indicating 700,000 KiB/s (that’s kibibytes per sec). That converts to 5734.4 Mbps. Considering I only have a 200 Mbps connection, I can only wish the reported numbers were true. According to pfSense (my router) I’m actually only transferring about 200Kbps inbound at the moment.

One thing I’ve noticed is that it looks like the sensors polling frequency changed from what it used to be, but maybe that’s just me misremembering.

In the meantime I divided values by 1000 getting something that looks fine …