release-23.1: sqlproxyccl: do not report BackendDown metrics on throttle and routing errors

[jaylim-crl]

Backport 2/2 commits from #108914.

/cc @cockroachdb/release

---

sqlproxyccl: do not report BackendDown metrics on throttle and routing errors

Previously, we were reporting the backend_down metric on the following errors:

• codeProxyRefusedConnection
• codeParamsRoutingFailed
• codeUnavailable

These errors do not imply that the backend is down. We originally introduced
this in #57431, but looking at the PR, it appears unintentional. This commit
fixes that by not reporting the backend_down metric when the proxy returns
such errors.

Release note: None

Epic: none

sqlproxyccl: rename codeBackendDown to codeBackendDialFailed

This commit renames codeBackendDown to codeBackendDialFailed to prevent
confusions by developers. Note that we don't rename the metric here to avoid
breaking downstream consumers. At the same time, we will remove the old
codeBackendRefusedTLS code as it does not serve any purpose, and there wasn't
a metric for it as well.

Release note: None

Epic: none

Release justification: This fixes accuracy issues with SQL Proxy metrics.

[blathers-crl]

Thanks for opening a backport.

Please check the backport criteria before merging:

• Patches should only be created for serious issues or test-only changes.
• 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 

[jeffswenson]

LGTM

[jaylim-crl]

TFTR!