[Supervisor] Slow loading when outside network

Tags: #<Tag:0x00007f326aa45da0>

Hi there!

I’m having issues currently that I didn’t have before with Supervisor or with addons. Whenever I go on my server from an external network on the Supervisor page or on a addon ingress page that I pinned on the side, it takes a lot of itme to load.

When I do the same but from the internal network, no issues. I don’t see any logs with issues.

Here is my setup:

  • RPI4 4GB RAM (using good power adapter)
  • 128Gb SANDISK usb hard drive (boot on usb)
  • Ubuntu 20.04 server LTS
  • Using docker image (followed Installing Home Assistant Supervised on Raspberry Pi OS)
  • Using duckdns + nginx
  • Conbee 2 attached
  • Addons: ADB, duckdns, file editor, grafana (disabled), home assistant google drive backup, influxdb (disabled), mariadb, mosquitto broker, nginx, zibee2mqtt edge

For info, the memory usage is about 26%, 13.6% of hard drive used, 35°C. In terms of internet connectivity, using Google DNS in my router, using IPV4 (I can"t seem to use V6 since I’m using a router behind the internet supplier router and passtrough just doesn’t work).
My download is 18.32Mb/s, upload 0.804Mb/s (using nPerf). When using my internet supplier status I’m at 20.7 and 1.1.

I tried to flash everything back from scratch but to no avail. I tried reinstalling every addon but doesn’t do anything either.

I really don’t know what to do, I hope I don’t need to use a NUC instead since the only issue is when I’m outside of the home network… Furthermore it’s only the supervisor, lovelace works fine, I can’t easily move between the menus et go into the HA config.

Using app on android or through chrome is the same (same on Windows using my phone connection).

Thanks for your help!

After checking, it seems like it is caused by /api/hassio/app/frontend_latest/entrypoint.6bc241a4.js that takes a lot of time to load. I made a performance test and it seems like it.

The guide is for Raspberry Pi OS, not Ubuntu on a Pi.

1 Like

Hi,

Thanks for the help, I switched and used the official install method on Debian 10 on the RPI4 and I have exactly the same issue with a 100% clean install of Debian + HA etc… I only have duckdns and nginx currently.

Sounds like a network configuration issue with Nginx. If it works locally and not remotely, something is not setup correctly.

It works remotely, it’s just super slow, only supervisor and apps, not HA itself. My guess is that the network between the docker images is super slow, maybe because of firewalls?

For NGINX, I didn’t change anything except the hostname:

domain: blablabla.duckdns.org
certfile: fullchain.pem
keyfile: privkey.pem
hsts: max-age=31536000; includeSubDomains
cloudflare: false
customize:
  active: false
  default: nginx_proxy_default*.conf
  servers: nginx_proxy/*.conf

I just tested behind my first router and same delay.

Just tested using HASSIO OS and same issue with a 100% clean install. Really strange… Going back to Debian. I’m almost sure that it’s caused by the entrypoint.js that takes too much time.

Having exactly the same issue. Drives me crazy on a regular base, trying different settings etc.

Glad to hear I am not alone. Just wondering, shouldn’t there be many more complaints from all type of users?

Hass OS user here btw.

Yeah I’m a bit suprised that no one has these issues, or maybe no one is accessing supervisor stuff from the outside network or maybe using a VPN?

Imo, there’s a firewall somewhere adding latency here.

Have you opened an issue?

Before opening an issue I wanted to see if other people from the community had this issue. Gonna open one as sonn as I have the time.