feat(analytics): implement dynamic cost and carbon factors with label-based attribution

[victoriacheng15]

Summary

Transformed the analytics service from static baselines to a dynamic, configuration-driven engine by externalizing energy cost and carbon intensity factors. Standardized feature attribution across the K3s infrastructure using consistent Kubernetes labels and updated Grafana dashboards to support real-time interactive parameter adjustments.

List of Changes

• Refactored ThanosResourceProvider to load CARBON_INTENSITY_G_KWH and ENERGY_COST_CAD_KWH from environment variables with safe fallbacks and startup logging.
• Enhanced analytics service attribution logic with expanded mappings for agentic infrastructure (MCP Hub) and standardized infrastructure tagging.
• Applied app.kubernetes.io/feature labels to all core K3s manifests (Postgres, Loki, Tempo, Thanos, Otel) to ensure consistent telemetry discovery.
• Decoupled Grafana dashboards from hardcoded constants by implementing template variables for cost and carbon factors, enabling interactive regional reporting.

Verification

• Verified analytics service logs factor loading on startup via unit tests and manual log inspection.
• Confirmed unit tests pass for the refactored analytics engine and service.
• Validated JSON schema compliance for updated Grafana dashboards and K3s manifests.