fix(curator): make the prod resource sizing branch reachable

[austinderrick]

The bug

curator.resources in _helpers.tpl picks a sizing block by environment when a release doesn't set resources itself:

{{- if eq .Values.environment "production" -}}

But values.schema.json restricts environment to dev | qa | prod, and the schema is enforced at render. "production" is a value the chart will not accept, so the production branch was unreachable — a prod release with no explicit resources silently got the small block (500m/512Mi requests and limits) instead of the intended 1 CPU / 1Gi requests and 1500m / 2Gi limits.

Blast radius: none today

All 19 prod-*-helmrelease.yaml files in iac-interworks (plus the .templates/ stub they're generated from) set resources explicitly, so every prod site today takes the override branch and never reaches this code. The fix corrects a latent default: it changes what a future prod site gets if it omits resources, and nothing about the current fleet.

The fix

• Compare against "prod", the value the schema actually permits.
• Note on the helper that the compared values have to stay in sync with the schema's environment enum, since that drift is what caused this.
• Add three cases to deployment_test.yaml: prod sizing, non-prod sizing, and an explicit resources override winning over both.

Testing

Ran the suite locally with the pinned helm 4.2.3 and helm-unittest v1.1.1:

• With the fix: 25 tests pass across all 6 suites.
• With the bug reintroduced (reverting only the comparison): the new prod-sizing case fails as it should — Expected to equal: 1 / Actual: 500m.

Found while writing REVIEW.md in #77, which cites this as the reference example of schema/template drift.