feat(platform): migrate hub to server2 and optimize for high performance

[victoriacheng15]

Summary

Migrated the entire Observability Hub stack from a resource-constrained laptop (Server 1) to a high-performance workstation (Server 2). Optimized the platform for large-scale telemetry workloads by doubling resource limits, resolving system-specific security constraints, and hardening service discovery.

List of Changes

• Infrastructure: Expanded root partition to 1TB and tuned K3s eviction thresholds to eliminate storage-induced pod evictions.
• Orchestration: Transitioned core services to LoadBalancer type to enable reliable localhost access for host-based agents and proxy.
• Performance: Implemented a 'High Performance' profile across Helm values, significantly increasing CPU/RAM ceilings for the LGTM stack.
• Security: Hardened the Fedora environment by applying targeted SELinux contexts to configuration files and binaries.
• Reliability: Provisioned missing S3 buckets and discovery services to restore real-time metrics and log ingestion pipelines.
• Code: Updated collectors to align with new host hardware (network device eno1) and modernized OpenBao initialization.

Verification

• Verified all 14 pods in observability namespace are in Running state.