Add proxy env variables to remote schema configuration

### Is your proposal related to a problem?

Yes, described in discussion https://github.com/hasura/graphql-engine/discussions/10242.

### Describe the solution you'd like

Add to remote schema configuration the following fields:
+ `HTTPS_PROXY`
+ `HTTP_PROXY`
+ `WSS_PROXY`
+ `WS_PROXY`

They control how hasura connects to a remote schema:
1/ `HTTPS_PROXY`, `HTTP_PROXY` for queries and mutations
2/ `WSS_PROXY`, `WS_PROXY` for subscriptions

Variables 1/ are typically used by [curl](https://everything.curl.dev/usingcurl/proxies/env.html), [wget](https://www.gnu.org/software/wget/manual/html_node/Proxies.html), Python [requests](https://www.zenrows.com/blog/python-requests-proxy#how-to).

Variables 2/ would be the same for `wss://` and `ws://` flows.
Though most corp proxies will not allow that traffic anyway it shouyld be possible to configure it.

If these values are not set, then no proxy is used for the connection to the remote schema.

The `NO_PROXY` field is not useful as the absence of `*_PROXY` has the same effect.

### Describe alternatives you've considered

Alternatively the fields above could be infered from the environment variables with the same name for all remote schemas.
But bad design, I think, as some remote schemas would be accessible through a proxy only, and other without only.
This is typically the case when an hasura graphql engine deployed on a corp premises has remote schemas inside (private) and outside (public) the corp network.

Start by reading the remote schema configuration and connection handling, using discussion #10242 for the intended behavior. Done means remote schemas can independently configure HTTPS_PROXY and HTTP_PROXY for queries and mutations, plus WSS_PROXY and WS_PROXY for subscriptions, with no proxy when unset.

Written by the indexing model from the issue text.