fix: CNS should not mark itself as ready if it can't write out a conflist.

[paulgmiller]

Reason for Change: CNS should not signal its healthz till it can write out a conflist. Going ready

1. makes it it harder for both engineers and systems to know CNS isn't getting to a good state.
2. makes it easier for a bad rollout to takea bug to all nodes.

Refactors

1. This moves alot of ncsync logic out of internalapi.go to synchostnc.go
2. It adds a new wait group like object (networkContainerSyncState) to httprestservice
3. Changes direct calls/loops in main to SyncHostNCVersion to use StartSyncHostNCVersionLoop for consistencey and better mangemetn of starting wait
4. has nodesubent also startwat/signal readiness and not write the conflist direcl
5. moves conf list writing to a wait function that also makes sure waits on nc (nodesubbnet and syncHostNCVersion) have signaled.

Tries to hide as much as possible to make movign to subpackage easier in future but pretty tangled with HttpRestServer still.

• [ ] uses conventional commit messages
• [ ] includes documentation
• [X] adds unit tests
• [ ] relevant PR labels added

Notes:

[rbtr]

/azp run Azure Container Networking PR

[azure-pipelines[bot]]

Azure Pipelines successfully started running 1 pipeline(s).

[github-actions[bot]]

This pull request is stale because it has been open for 2 weeks with no activity. Remove stale label or comment or this will be closed in 7 days

[github-actions[bot]]

Pull request closed due to inactivity.