fix: avoid metric increments on no-op delete reconcile paths

[felipesabadini]

What type of PR is this?

fix: avoid metric increments on no-op delete reconcile paths

What this PR does / why we need it

This PR fixes metric inflation during no-op reconciles for optional resources.

Previously, deleteDeployment, deleteDaemonSet, deleteHPA, and deletePDB always called DeleteAllOf and recorded resource_delete_* success metrics, even when no matching resources existed. In default Deployment mode (without DaemonSet/HPA/PDB), this caused delete metrics to grow continuously despite no real deletions.

Additionally, createOrUpdateHPA and createOrUpdatePodDisruptionBudget registered apply metric defers before nil checks, causing resource_apply_total to be incremented on the nil->delete path.

Changes

• Add pre-delete existence checks (List with the same namespace/label selector) in:
  • deleteDeployment
  • deleteDaemonSet
  • deleteHPA
  • deletePDB
• If no matching resources exist, return early and skip delete metric recording.
• Move apply-metric defer blocks in:
  • createOrUpdateHPA
  • createOrUpdatePodDisruptionBudget
    to run only in actual apply paths (after nil checks), matching Deployment/DaemonSet behavior.

Which issue(s) this PR fixes

Fixes #8438

Release Notes

No

[netlify]

✅ Deploy Preview for cerulean-figolla-1f9435 ready!

Name	Link
🔨 Latest commit	c528d5b
🔍 Latest deploy log	https://app.netlify.com/projects/cerulean-figolla-1f9435/deploys/69cbf9126b753700080fa964
😎 Deploy Preview	https://deploy-preview-8480--cerulean-figolla-1f9435.netlify.app
📱 Preview on mobile	Toggle QR Code... Use your smartphone camera to open QR code link.

---

To edit notification comments on pull requests, go to your Netlify project configuration.

[felipesabadini]

Validated in a local Kind test cluster (default Deployment mode, no HPA/PDB/DaemonSet configured).

• Deployed Envoy Gateway with this fix.
• Checked /metrics after startup and again after ~60s of reconcile loops.
• Re-checked across multiple reconcile cycles.

Observed:

• resource_delete_total: no increments for DaemonSet, HPA, or PDB (previously these grew continuously on no-op reconciles).
• resource_apply_total: no increments for HPA/PDB on nil -> delete paths (previously incremented incorrectly).
• Counters remained stable over time.

[codecov]

Codecov Report

❌ Patch coverage is 70.96774% with 18 lines in your changes missing coverage. Please review.
✅ Project coverage is 74.34%. Comparing base (c86eb78) to head (c528d5b).
⚠️ Report is 18 commits behind head on main.

Files with missing lines	Patch %	Lines
...ternal/infrastructure/kubernetes/infra_resource.go	70.96%	10 Missing and 8 partials ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##             main    #8480      +/-   ##
==========================================
- Coverage   74.38%   74.34%   -0.05%     
==========================================
  Files         243      243              
  Lines       38227    38262      +35     
==========================================
+ Hits        28436    28444       +8     
- Misses       7802     7823      +21     
- Partials     1989     1995       +6     

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

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.
• 📦 JS Bundle Analysis: Save yourself from yourself by tracking and limiting bundle sizes in JS merges.

[kkk777-7]

Hi @felipesabadini, thanks for working this!
Overall LGTM, could you fix lint?

[jukie]

This will fix linting issues:

[jukie]

/retest

[arkodg]

hey @felipesabadini can you help with these follow ups

• createOrUpdateHPA() now returns directly on r.HorizontalPodAutoscaler()
  errors before the defer is installed, so HPA render/patch failures no longer
  increment resource_apply_total{result="failure"}.
• deleteDeployment(), deleteDaemonSet(), deleteHPA(), and deletePDB()
  now return on the pre-delete List() call before the delete-metrics defer is
  installed, so list/RBAC/API failures no longer increment
  resource_delete_total{result="failure"}.

#8438 looks fixed for the no-op delete case, but failure-path observability
regressed