Hi, sorry for the big delay responding to your question!
Can I clarify one thing: based on the current configuration, I would expect a timeout of 65 seconds. Is your app showing you the number 60 somewhere, or did you time it?
Regardless of that, I think I am going slightly mad as I believed I had already set the nginx reverse proxy's keepalive_timeout to 0, thereby disabling it entirely, allowing any such timeouts to be handled per-stack if required.
Would a default keepalive_timeout 0 in the nginx reverse proxy config resolve your use-case?
Otherwise (and in the meantime), if you'd like to override it yourself, it's certainly possible, although currently deliberately undocumented. See here for more info.