Fix controller memory usage

[zzjin]

Decrease 50x footprint by using filtered cache. fixes #4588.

When deploy controller to clusters witch have large amount of resources like deployments/pods/secrects/ingresses, some controller is consume more memory than needed.

Before	After

---

Key problem:

When controller manager start, it will watch and cache all needed resource it owns for performance reason.
When cluster have thousand of resources, the cache memory increases whenever custom resource's count.

Solution:

With terminal/adminer controller usage, controller manager only want to watch resources we created and/or managed for update/delete etc events.
So add labels to controller managed resources and then create cache filtered only for those labeled resources, can reduce cache memory usage to match real custom resource's count, and also reduce api-server call response to save internal network bandwidth.

[pull-request-size]

Whoa! Easy there, Partner!

This PR is too big. Please break it up into smaller PRs.

[sealos-ci-robot]

🤖 Generated by lychee action

Summary

Status	Count
🔍 Total	1198
✅ Successful	371
⏳ Timeouts	0
🔀 Redirected	0
👻 Excluded	826
❓ Unknown	0
🚫 Errors	0

Full action output

Full Github Actions output

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 65.45%. Comparing base (cbbe279) to head (575c67e).

Additional details and impacted files

@@           Coverage Diff           @@
##             main    #4589   +/-   ##
=======================================
  Coverage   65.45%   65.45%           
=======================================
  Files           8        8           
  Lines         660      660           
=======================================
  Hits          432      432           
  Misses        180      180           
  Partials       48       48           

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.