I have recently been having a lot of issues maintaining a connection with my proton vpn in gluetun. it keeps dropping and reconnecting due to failing periodic small health check. My internet connection is solid without connection issues, running a symmetrical 1gbit and the router (gl.inet brume 2) doesn’t register any dropouts in connection.

i have tried following the suggested troubleshooting in the docs, updating the vpn server list etc., but none of the suggestions seem to fix this issue. it has been going on for several months now. if i use mullvad it works perfectly without connection issues, but this is destroying my seed ratio due to no port forwarding.

docker compose

log from container

  • Appoxo@lemmy.dbzer0.comEnglish
    1·
    6 days ago

    I use gluetun with my own ovpn file. I have no knowledge on other vpn-integrations!

    I see some issues with the provided info

    There are some ping issues in the log
    Have you a monitor service like uptime kuma that you can cross-check?
    Best case to be configurable to the frequency gluetun checks for

    2026-03-15T07:04:49Z INFO [vpn] 👉 See https://github.com/qdm12/gluetun-wiki/blob/main/faq/healthcheck.md

    Have you tried that before posting?

    2026-03-15T07:04:50Z INFO [openvpn] UDPv4 link remote: [AF_INET]46.29.25.4:1194

    2026-03-15T07:05:50Z WARN [openvpn] TLS Error: TLS key negotiation failed to occur within 60 seconds (check your network connectivity)

    🚒🚒🚒🚒🚒🚨🚨🚨🚨🚨🚨🚒🚒🚒🚒🚒 (…)

    That seems bad.

    Side quests:

    • Have you a separate system or VM that you could test your config on a fresh vanilla system? (to rule out issues related to networking tweaks)
    • Have you the issue with other connections?
    • Do you notice a pattern? Maybe in time (clock), duration of connection (how long did it work before the issue arrised? Does Proton force kill connections periodically?) Maybe a program/service executing in the background that touches the network config?
    • When was the last config change? Yesterday? Two days ago? (Example: My traefik cert failed because I fucked with the config. I also touched my 2FA proxy container redirect which resulted in a faulty integration and me being unable to login. The time since I have touched the config was 3 days prior.)
    • ExcessShiv@lemmy.dbzer0.comOPEnglish
      1·
      6 days ago

      It worked without issues for about 1.5year without me changing anything in the compose file, about 6 months ago I started noticing health issues with the container. My torrent stack has been untouched since I set it up 2 years ago.

      There doesn’t seem to be a clear pattern in failures, sometimes its 20min, sometimes its 8 hours between failures.

      • Venia Silente@lemmy.dbzer0.comEnglish
        1·
        6 days ago

        Has your gluetun image been kept up-to-date, btw? Changes to compose file or not, there were IIRC two breaking changes in gluetun connectivity with Proton due to changes in the Proton CLI, the two of them were recent enough, I think in v3.38 and v3.40.

        • ExcessShiv@lemmy.dbzer0.comOPEnglish
          1·
          6 days ago

          It is updated automatically with watchtower, but I have manually pulled the latest a few times, most recently about two months ago.

          Edit: weirdly enough, despite me pulling the latest image just now, which was released two days ago, the log still says I’m running 4 commits behind when I start the container 🤔

          • Appoxo@lemmy.dbzer0.comEnglish
            2·
            6 days ago

            Don’t know if proton does offer it, but try to run the raw openvpn (or what the format of proton uses) config in gluetun.
            Maybe theres an issue with the native integration.

            If it runs fine, I’d say it’s a bug
            If it runs just as bad, theres a system issue.

          • blueduck@piefed.socialEnglish
            1·
            6 days ago

            I ran gluetun + proton for 4-5 years. v3.41 broke the integration completely and I could never get it working. Tried both wireguard and ovpn setups.

            My best advice is pin to the 3.40 release or change providers. The maintainer hasn’t address the specific problem with proton in over a year. There was some discussion recently about improvements to the healthcheck endpoint, so he is working on the project. But it’s slow going and proton is the only provider I know about that is having consistent issues