Catch exception when subscribing to NetworkChange

[rzikm]

Fixes #94794.

We can simply swallow the exception and not monitor network changes. RFC allows us to.

RFC: https://tools.ietf.org/html/rfc7838#section-2.2

When alternative services are used to send a client to the most
optimal server, a change in network configuration can result in
cached values becoming suboptimal. Therefore, clients SHOULD remove
from cache all alternative services that lack the "persist" flag with
the value "1" when they detect such a change, when information about
network state is available.

[ghost]

Tagging subscribers to this area: @dotnet/ncl
See info in area-owners.md if you want to be subscribed.

Issue Details

---

Fixes #94794.

We can simply swallow the exception and not monitor network changes. RFC allows us to.

RFC: https://tools.ietf.org/html/rfc7838#section-2.2

When alternative services are used to send a client to the most
optimal server, a change in network configuration can result in
cached values becoming suboptimal. Therefore, clients SHOULD remove
from cache all alternative services that lack the "persist" flag with
the value "1" when they detect such a change, when information about
network state is available.

Author:	rzikm
Assignees:	-
Labels:	area-System.Net.Http
Milestone:	-

[ManickaP]

LGTM, thanks.

If we're gonna backport it, we should test it (including the cleanup code). Even if just manually with faking the exception for the sake of the test.

[rzikm]

CI failures are unrelated

[rzikm]

/backport to release/8.0-staging

[github-actions]

Started backporting to release/8.0-staging: https://github.com/dotnet/runtime/actions/runs/6958490464