[Serve] integrate with Ray structured logging

[GeneDer]

Why are these changes needed?

Integrated Ray structured logging into Serve's logger. Two things are done in this PR:

1. Moved the logics to flatten serve's special key for extra logging attributes to _append_flatten_attributes() and called it from generate_record_format_attrs() so Ray structured logging can apply those nested attributes
2. Refactored Serve's json formatter into context filters and applied them onto Serve's log handlers. Reuses core's JSONFormatter when formatting structured logs

The main difference after integrated with Ray structured logging is there are new attributes provided by Ray (e.g. node_id, filename, and lineno). Also, the message format is just the message body itself, without filename and line number prefix.

Example log before the change

{
   "levelname":"INFO",
   "asctime":"2024-06-24 22:15:34,526",
   "actor_id":"59ef544fa6127fffd1b6c4ad01000000",
   "worker_id":"ce0efd00c47b08f51a5c04e2cdfb8919ed6df4d0b4b938d3e75b4321",
   "deployment":"app_Counter",
   "replica":"j8fhh1bm",
   "component_name":"replica",
   "request_id":"6f04a2fd-db85-4687-8302-ef478e7abee8",
   "route":"/favicon.ico",
   "application":"app",
   "message":"replica.py:373 - __CALL__ OK 2.8ms"
}

Example log after the change

{
   "asctime":"2024-06-24 22:10:57,906",
   "levelname":"INFO",
   "message":"__CALL__ OK 4.3ms",
   "filename":"replica.py",
   "lineno":373,
   "worker_id":"833490a0fcc662b86c04403694c85b1d14364a05b5c6d6dfec5d5d47",
   "node_id":"1f3b2424bccc7f407fad2bf549954b723ea46405dded64d7386239a7",
   "actor_id":"b5f78e9af7148f019d9e89ab01000000",
   "route":"/favicon.ico",
   "request_id":"59f62350-7da6-47a2-a964-b6dd4e512003",
   "application":"app",
   "deployment":"app_Counter",
   "replica":"8p7guoov",
   "component_name":"replica"
}

Related issue number

Closes #46125

Checks

• I've signed off every commit(by using the -s flag, i.e., git commit -s) in this PR.
• I've run scripts/format.sh to lint the changes in this PR.
• I've included any doc changes needed for https://docs.ray.io/en/master/.
  • I've added any new APIs to the API Reference. For example, if I added a
    method in Tune, I've added it in doc/source/tune/api/ under the
    corresponding .rst file.
• I've made sure the tests are passing. Note that there might be a few flaky tests, see the recent failures at https://flakey-tests.ray.io/
• Testing Strategy
  • Unit tests
  • Release tests
  • This PR is not tested :(

[edoakes]

As discussed offline let's pop off task_id and job_id for serve replica logs.

Also why is this now included? "serve_access_log":true

[GeneDer]

As discussed offline let's pop off task_id and job_id for serve replica logs.

Also why is this now included? "serve_access_log":true

oops, didn't pay attention, will remove serve_access_log. But basically core's formatter will take all the attributes on the log record

[jjyao]

So we decided to only config "ray.serve" logger not the root logger?

[GeneDer]

So we decided to only config "ray.serve" logger not the root logger?

Yea, I still feel it's a bit odd (and unnecessary) for Serve to configure the root logger. We can follow up on this if there are users who wants it :)