Uploads >60s over HTTP/2 are terminated by Traefik due to readTimeout

[tom-0221]

Uploads that take longer than 60 seconds and use HTTP/2 are terminated by Traefik due to the default readTimeout of 60 seconds. This issue does not occur when using HTTP/3.

How to reproduce:

1. Open Firefox and go to about:config
2. Set network.http.http3.enabled to false to force HTTP/2
3. Start a file upload via the Nextcloud web interface (Use a tool like NetLimiter to throttle the upload speed so that it lasts longer than 60 seconds)
4. The upload fails (with a 502 Bad Gateway Timeout)

Root cause:
Traefik’s default readTimeout is set to 60 seconds. Under HTTP/2, if no data is received during that period, Traefik closes the connection—even if the upload is still in progress.

Workaround:
Setting this label on the affected container resolves the issue:

traefik.http.services.nextcloud.loadbalancer.server.readtimeout=120s

Suggestion:
It would be helpful if mash could either:

• Document this edge case for Nextcloud and similar apps with long-lived uploads
• Or apply a higher default readTimeout for such cases automatically

[hapm]

This could be related to #372

[dasTholo]

The mash-nextcloud service has not started with this label.
Yes i test with "traefik.http.services.mash-nextcloud.loadbalancer.server.readtimeout=120s"
With Global i can upload my backups from HomeAssistent with Webdav sucessfully.
traefik_config_entrypoint_web_secure_transport_respondingTimeouts_readTimeout: 30m

Maybe mash must impliment a servertransport file like this