release-21.2: rpc: avoid network IO in Dialer.ConnHealth

[blathers-crl]

Backport 1/1 commits from #70017 on behalf of @erikgrinaker.

/cc @cockroachdb/release

---

Dialer.ConnHealth is used to check whether a healthy RPC connection
exists to a given node, in order to avoid interacting with unavailable
nodes. However, this actually attempted to dial the node if no
connection was found, which can block for tens of seconds in the case of
an unresponsive node. This is problematic since it is used in
performance-critical code paths, including Raft command application.

This patch changes Dialer.ConnHealth to avoid dialing the node, and
adds a Context.ConnHealth helper with access to the RPC connection
registry. Because DistSQLPlanner relied on ConnHealth to dial the
remote node, it also adds Dialer.ConnHealthTryDial which retains
the old behavior for use in DistSQL until a better solution can be
implemented.

Resolves #69888.

Release note (bug fix): Avoid dialing nodes in performance-critical code
paths, which could cause substantial latency when encountering
unresponsive nodes (e.g. when a VM or server is shut down).

---

Release justification:

[blathers-crl]

Thanks for opening a backport.

Please check the backport criteria before merging:

• Patches should only be created for serious issues.
• Patches should not break backwards-compatibility.
• Patches should change as little code as possible.
• Patches should not change on-disk formats or node communication protocols.
• Patches should not add new functionality.
• Patches must not add, edit, or otherwise modify cluster versions; or add version gates.

If some of the basic criteria cannot be satisfied, ensure that the exceptional criteria are satisfied within.

• There is a high priority need for the functionality that cannot wait until the next release and is difficult to address in another way.
• The new functionality is additive-only and only runs for clusters which have specifically “opted in” to it (e.g. by a cluster setting).
• New code is protected by a conditional check that is trivial to verify and ensures that it only runs for opt-in clusters.
• The PM and TL on the team that owns the changed code have signed off that the change obeys the above rules.

Add a brief release justification to the body of your PR to justify this backport.

Some other things to consider:

• What did we do to ensure that a user that doesn’t know & care about this backport, has no idea that it happened?
• Will this work in a cluster of mixed patch versions? Did we test that?
• If a user upgrades a patch version, uses this feature, and then downgrades, what happens?

[cockroach-teamcity]

This change is