chore: bump flask backend resource limits (CPU 0.5, mem 1Gi)

What

Increase monitoring_flask_backend container resources for future deployments.

Docker Compose (docker-compose.yml)

	Before	After
CPU limit	0.1 cores	0.5 cores
Memory limit	192MiB (192m)	1GiB (1073741824 bytes)

Helm (postgres_ai_helm/values.yaml)

Previously resources: {} (no requests or limits). Now:

• Requests: CPU 500m, memory 256Mi
• Limits: CPU 500m, memory 1Gi

Why

Flask backend was the most resource-starved service in the monitoring stack. Under real customer load, the old 0.1-core CPU cap and 192MiB memory cap can cause CPU throttling, request timeouts, and OOM kills.

The Helm CPU request now matches the 500m CPU limit so Kubernetes schedules the backend with the capacity this MR expects instead of reserving the old 0.1-core baseline under node contention. The memory request remains conservative at 256Mi to avoid over-reserving cluster memory by default, while the 1Gi limit gives the backend headroom before OOMKill.

Out of scope: HPA/VPA and environment-specific overrides; this MR only raises the default Docker Compose and Helm resources.

Validation

• ruby -e 'require "yaml"; ...' resource assertions for Compose and Helm values
• git diff --check origin/main...HEAD
• PGAI_TAG=dev docker compose -f docker-compose.yml config --quiet
• python3 -m pytest tests/compliance_vectors/test_flask_resources.py tests/compliance_vectors/test_vm_auth.py -q (14 passed, 11 skipped locally because helm is not installed)
• CI vm-auth:config:tests now runs helm lint, helm template, and the Flask resource render test with Helm installed

Closes #168 (closed) Related: #130 Requested by @dfomin.