Add new field to health status

[rsajwani]

Description

This PR introduce a new boolean in HealthStatus structure. The idea is to distinguish between Vtorc being up and running as service and Vtorc being ready to run recoveries after having run the discovery check. The way code has been setup when you hit /debug/health,

1. Healthy flag will tell you if the service is up and running.
2. HasDiscoveredOnce flag will tell you if Vtorc has completed the initial round of discovery. This will ensure that we don't run into situations as pointed out in Feature Request: VTOrc should be "ready" only when its ready to remediate. #12268.

We only return a 200 status when VTOrc is Healthy and has run the first round of discovery.

Solution

While resolving this issue, I thought there can be multiple ways we can implement it.

1. We could have created a separate endpoint e.g /debug/discoveryHealth. But I thought it would be overkill and given we already are exposing HealthStatus through health endpoint, why not just add an extra field there.
2. Inside ContinuousDiscovery, before the for {...} we can set boolean process.HitAtLeastOneDiscovery =true in a separate go routine by just waiting for TopoInformationRefreshSeconds. E.g

go func() {
		time.Sleep(config.Config.TopoInformationRefreshSeconds * time.Second)
		process.HitAtLeastOneDiscovery.CompareAndSwap(false, true)
	}()

This will run only once.

3. Inside DiscoverInstance we set process.HitAtLeastOneDiscovery =true. This will ensure that we are flagging only when we have at-least one discovery. But problem with this approach is it will do memory read of boolean (atomic so will be very fast and light) during every successful DiscoverInstance call. Also, doing just one discovery isn't enough, we need to wait for all the discoveries to complete.

4. Add a wait group in the topoinformation tick and wait for the refreshes to complete. Set a boolean value when the refreshes are complete. Use this boolean in the health field.

I went with option 4 but I am open to discuss above mentioned solution or any other one not listed here.

Related Issue(s)

closes #12268

Checklist

• "Backport to:" labels have been added if this change should be back-ported
• Tests were added or are not required
• Did the new or modified tests pass consistently locally and on the CI
• Documentation was added or is not required

Deployment Notes

[vitess-bot]

Review Checklist

Hello reviewers! 👋 Please follow this checklist when reviewing this Pull Request.

General

• Ensure that the Pull Request has a descriptive title.
• If this is a change that users need to know about, please apply the release notes (needs details) label so that merging is blocked unless the summary release notes document is included.
• If a test is added or modified, there should be a documentation on top of the test to explain what the expected behavior is what the test does.

If a new flag is being introduced:

• Is it really necessary to add this flag?
• Flag names should be clear and intuitive (as far as possible)
• Help text should be descriptive.
• Flag names should use dashes (-) as word separators rather than underscores (_).

If a workflow is added or modified:

• Each item in Jobs should be named in order to mark it as required.
• If the workflow should be required, the maintainer team should be notified.

Bug fixes

• There should be at least one unit or end-to-end test.
• The Pull Request description should include a link to an issue that describes the bug.

Non-trivial changes

• There should be some code comments as to why things are implemented the way they are.

New/Existing features

• Should be documented, either by modifying the existing documentation or creating new documentation.
• New features should have a link to a feature request issue or an RFC that documents the use cases, corner cases and test cases.

Backward compatibility

• Protobuf changes should be wire-compatible.
• Changes to _vt tables and RPCs need to be backward compatible.
• vtctl command output order should be stable and awk-able.
• RPC changes should be compatible with vitess-operator
• If a flag is removed, then it should also be removed from VTop, if used there.

[rsajwani]

@deepthi @GuptaManan100 , should we add this in release notes?

[rsajwani]

I find out not every time Vtorc health is ready right away. So I am calling retry here. But retry was having assert inside it implementation. This force me to pull assert out of that method.

[rsajwani]

only after first successful discovery. Should we instead make it only after first discoveryInstance call regardless of success or failure?
I though successful is better.

[deepthi]

@deepthi @GuptaManan100 , should we add this in release notes?

Not needed

[deepthi]

Under what conditions does /debug/health now return a 200 (OK) response? That should be stated clearly in the PR description itself.

[GuptaManan100]

There are 2 changes that are required -

1. We actually want to only mark VTOrc as healthy after it has read all the instances for the first time. I would mark the boolean true in case <-tabletTopoTick: of ContinuousDiscovery after waiting for both the refreshes to complete.
2. The return status of the API is very important. Its code should be 200 only after the discovery has completed.

[GuptaManan100]

I have made the requested changes 👍

[GuptaManan100]

I have updated the tests and description too. Please take a look @deepthi.

[deepthi]

A couple of naming suggestions, the rest LGTM

[shlomi-noach]

Looks good

[shlomi-noach]

Looks good