fix: increase production frontend memory limits to prevent OOMKills

- Frontend memory: 512Mi → 2Gi limit, 256Mi → 512Mi request
- OAuth proxy limits: 256Mi request, 512Mi limit (NEW - prevent unbounded growth)

Fixes sawtooth memory pattern causing pod restarts and connection errors.
Matches LiteLLM precedent (commit b43bf39: 512Mi → 2Gi, 4x increase).

Production only - other environments not affected by OOMKills.

Co-Authored-By: Claude Sonnet 4.5 <[email protected]>

Author: jeremyeder

components/manifests/overlays/production/frontend-oauth-deployment-patch.yaml

@@ -1,4 +1,6 @@
-# Patch to add OAuth proxy sidecar to frontend deployment
+# Patch for production frontend deployment
+# - Adds OAuth proxy sidecar for authentication
+# - Overrides resource limits to prevent OOMKills (sawtooth memory pattern)
 apiVersion: apps/v1
 kind: Deployment
 metadata:
@@ -7,6 +9,15 @@ spec:
   template:
     spec:
       containers:
+      # Frontend container - increased limits to prevent OOMKills
+      - name: frontend
+        resources:
+          requests:
+            memory: 512Mi
+            cpu: 200m
+          limits:
+            memory: 2Gi
+            cpu: 1000m
       # OAuth proxy sidecar
       - name: oauth-proxy
         image: quay.io/openshift/origin-oauth-proxy:4.14
@@ -46,6 +57,13 @@ spec:
           periodSeconds: 5
           successThreshold: 1
           failureThreshold: 3
+        resources:
+          requests:
+            memory: 256Mi
+            cpu: 50m
+          limits:
+            memory: 512Mi
+            cpu: 200m
         volumeMounts:
         - mountPath: /etc/oauth/config
           name: oauth-config