fix(healthcheck): Update Hubble Readiness & Liveliness probes

[mereta]

Description

Updating Readiness probe to use hubble status command
Adding a liveliness probe to use hubble status command

Related Issue

#1047

Checklist

• I have read the contributing documentation.
• I signed and signed-off the commits (git commit -S -s ...). See this documentation on signing commits.
• I have correctly attributed the author(s) of the code.
• I have tested the changes locally.
• I have followed the project's style guidelines.
• I have updated the documentation, if necessary.
• I have added tests, if applicable.

Screenshots (if applicable) or Testing Completed

Checks functioning as expected. Tested with an invalid tcp address to make the gRPC connection for hubble status and containers get recycled as expected.

Additional Notes

---

[matmerr]

How long does the hubble status cmd take to respond at scale? If the command is generally quick and fits within the 30s window, perfect, but if it's frequently greater than 30s for a healthy response then restarting the pod may do more harm than good

[mereta]

Going to run some scale tests. TBC.

[anubhabMajumdar]

Ref: https://github.com/cilium/cilium/blob/c7347199af4fe426c87fc0f5572f493559259ea6/pkg/hubble/observer/local_observer.go#L210

The status should be pretty static w.r.t. scale. Should have the ring buffer size (which is fixed) as upper limit.

[mereta]

Also some adjustments needed to run scale tests for hubble atm

[MikeZappa87]

Do any conditions exist with Hubble status where the exit code is 0 however the retina process itself is in a bad state?

[anubhabMajumdar]

Do any conditions exist with Hubble status where the exit code is 0 however the retina process itself is in a bad state?

Following the code of Hubble status here I think there's no such scenario if hubble status -ocompact is not used

[MikeZappa87]

Do any conditions exist with Hubble status where the exit code is 0 however the retina process itself is in a bad state?

Following the code of Hubble status here I think there's no such scenario if hubble status -ocompact is not used

If retina is deadlocked how does the current approach determine this?