[Serve] [Dashboard] Add serve controller metrics to serve system dashboard page

[architkulkarni]

Why are these changes needed?

Adds the following Serve controller metrics to the Serve details page on the dashboard:

ray_serve_num_ongoing_http_requests
ray_serve_num_ongoing_grpc_requests
ray_serve_controller_num_starts
ray_serve_num_scheduling_tasks
ray_serve_num_scheduling_tasks_in_backoff
ray_serve_controller_num_control_loops

#43797 (comment) contains a screenshot.

Related issue number

Closes #35667

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 :(

[shrekris-anyscale]

Awesome, thanks for getting this out so quickly! The dashboard page looks good.

It's a little odd that ray_serve_controller_num_starts, ray_serve_num_scheduling_tasks, and ray_serve_num_scheduling_tasks_in_backoff aren't showing up. ray_serve_controller_num_starts should show up soon after the controller initializes for the first time. ray_serve_num_scheduling_tasks and ray_serve_num_scheduling_tasks_in_backoff should show up after a few requests on a hello world app.

Is there any chance that the query is filtering for an incorrect/unexpected tag?

[GeneDer]

I think ray_serve_num_scheduling_tasks, and ray_serve_num_scheduling_tasks_in_backoff are both gauge just happened to set to 0. There actually data in your screenshot :)

ray_serve_controller_num_starts is indeed the odd one that doesn't come with data. Would need to dig into how come it didn't log a metrics on start. Could be a bug in the code somewhere 😕

[architkulkarni]

Thanks for the reviews, I'll update the PR soon.

@edoakes any thoughts on other metrics that should be included here?

[edoakes]

Thanks for the reviews, I'll update the PR soon.

@edoakes any thoughts on other metrics that should be included here?

Oh nice, this is looking good! Could we please also add serve_controller_num_control_loops? This detects if the controller is "stuck" (counter is not increasing)

[architkulkarni]

Will remove serve_controller_num_starts for now since it doesn't seem like it's being exported by Serve. Created an issue here: #43829

[edoakes]

Will remove serve_controller_num_starts for now since it doesn't seem like it's being exported by Serve. Created an issue here: #43829

NP. This is somewhat redundant with the control loop counter anyways

[GeneDer]

LGTM!

[architkulkarni]

Latest commit looks good, merging