Some of our users have reported the ssh to cluster would suffer delay or disconnecting sometimes. This PR adds more metrics and logs to help clarify the different disconnection reasons:

• ClientClosed, which is good, note that if the server PING is not receive a PONG in time (20s timeout by default), it is also considered as a ClientDisconnect. The PING/PONG might be delayed by the event loop issue, but I never produce a case that long event loop blocks cause PING/PONG timeout.
• SSHToPodDisconnected, the ssh connection to pod is disconnected, usually caused by pod issue, e.g. the Pod is oomkilled
• KubectlPortForwardExit, the kubectl port-forward process under the ssh connection exits, I can reproduce this with very low probability when I start many long-lived ssh connections. The reason why the port-forward process exit is a follow-up and can be prioritized if we can have non-trivial probability of repros in any environment.

The 3 cases are manually tested via killing port-forward process / pod / ssh process manually.

Tested (run the relevant ones):

• Code formatting: install pre-commit (auto-check on commit) or bash format.sh
• Any manual or new tests for this PR (please specify below)
• All smoke tests: /smoke-test (CI) or pytest tests/test_smoke.py (local)
• Relevant individual tests: /smoke-test -k test_name (CI) or pytest tests/test_smoke.py::test_name (local)
• Backward compatibility: /quicktest-core (CI) or pytest tests/smoke_tests/test_backward_compat.py (local)