release-22.1: ccl/sqlproxyccl: invoke rebalancing logic during RUNNING pod events

[blathers-crl]

Backport 2/2 commits from #81177 on behalf of @jaylim-crl.

/cc @cockroachdb/release

---

ccl/sqlproxyccl: invoke rebalancing logic during RUNNING pod events

This commit invokes the rebalancing logic during RUNNING pod events as part of
the pod watcher. Since the rebalancing logic depends on the tenant directory,
the pod watcher will now only emit events once the directory has been updated.
This is done for better responsiveness, i.e. the moment a new SQL pod gets
added, we would like to rebalance all connections to the tenant.

Note that the Watch endpoint on the tenant directory server currently emits
events in multiple cases: changes to load, and changes to pod (added/modified/
deleted). The plan is to update the tenant directory server to only emit events
for pod updates. The next commit will rate limit the number of times the
rebalancing logic for a given tenant can be called.

At the same time, we introduce a new test static directory server which does
not automatically spin up tenants for us (i.e. SQL pods for tenants can now
be managed manually, giving more control to tests).

ccl/sqlproxyccl: rate limit the number of rebalances per tenant

This commit rate limits the number of rebalances per tenant to once every
15 seconds (i.e. 1/2 of the rebalance loop interval). The main purpose of
this is to prevent a burst of pod events for the same tenant causing multiple
rebalances, which may move a lot of connections around.

Release note: None

---

Release justification: sqlproxy only change.

[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!