requests-proxy deployment hardcodes resources instead of reading from values

[saadqbal]

Problem

client/templates/requests-proxy-deployment.yaml hardcodes its container resources:

resources:
  requests:
    cpu: 100m
    memory: 256Mi
  limits:
    cpu: 1000m
    memory: 512Mi

Every other component (jobsManager, podsMonitor, mysql) reads from .Values.resources.<component>.* with defaults. The schema (client/values.schema.json) even defines resources.requestsProxy with requests / limits validation, but no template references it — so operator overrides to resources.requestsProxy.requests.cpu are silently ignored.

Fix

1. Add a resources.requestsProxy block in client/values.yaml (mirroring jobsManager / podsMonitor) with the current hardcoded values as defaults.
2. Update the template to read from .Values.resources.requestsProxy.* with | default fallbacks matching existing component patterns.

Acceptance criteria

• Default rendered values exactly match the current hardcoded values (no behavior change on default installs).
• --set resources.requestsProxy.limits.memory=1Gi is honored.
• helm-unittest covers the override path.

Severity

Low. Functional cluster works with hardcoded values; this is hygiene to match the other components and unblock operator-side tuning.

Found by

cursor[bot] on PR #137 review, 2026-05-20.