Still tinkering and trying to figure things out.
Some more links (to complement the first set)
- the ngx_http_proxy_module
- Nginx proxy_read_timeout vs. proxy_connect_timeout
- Issue with nginx reverse proxy
- why is nginx timeout-ing?
So the problem isn’t – as hoped as of the previous update – that the issue was the https-specific config.
In the latest attempt, per the GH discussion linked above, I’ve added a redirect rule to the nginx subdomain config file so that /radicale is pointed to /radicale/. This seems to have done nothing, other than prettify the file a bit.
Next attempt: adjusting the timeouts.