[RaphaelFakhri]

What this PR does / why we need it:

The PodDisruptionBudget in the kong and kong-ai-gateway charts built its selector from kong.metaLabels, which includes helm.sh/chart and app.kubernetes.io/version. An upgrade that changes either label updates the PDB selector immediately, while the running pods keep the old values. Until the rollout replaces them, the PDB covers none of those pods, so a node drain can evict all of them at once.

This PR switches the PDB selector to kong.selectorLabels, the same labels the Deployment uses for its own selector (app.kubernetes.io/name, app.kubernetes.io/component, app.kubernetes.io/instance). The PDB metadata labels are unchanged.

• charts/kong/templates/pdb.yaml, charts/kong-ai-gateway/templates/pdb.yaml: use kong.selectorLabels in spec.selector.matchLabels.
• charts/kong/ci/__snapshots__/pdb-default.snap, pdb-always-allow.snap: rebuilt with helm chartsnap -u.
• CHANGELOG entries for both charts.

Which issue this PR fixes

• fixes kong: PodDisruptionBudget selector includes chart and version labels, so it stops matching running pods after an upgrade #1592

Special notes for your reviewer:

policy/v1 PodDisruptionBudget selectors are mutable, so helm upgrade updates existing PDBs in place.

make test.golden equivalent (helm chartsnap -c ./charts/kong -f ./charts/kong/ci/ with the Makefile's --api-versions) passes: 37 snapshots matched. helm lint passes for both charts.

Checklist

• PR is based off the current tip of the main branch.
• Changes are documented under the "Unreleased" header in CHANGELOG.md
• Commits follow the Kong commit message guidelines