[Feature] Autonomous System Health Inspection via Hera Agent

### Background
Static thresholds often lead to "alert fatigue" or missed silent failures. We need a more proactive way to inspect system health across Metrics, Tracing, and Logging.

### Proposed Feature
Implement an AI-driven inspection worker:
- **Scheduled Inspection**: Hera Agent periodically pulls data from Prometheus and Jaeger.
- **Anomaly Detection**: Instead of fixed thresholds, use AI to detect "weird" patterns (e.g., a sudden drop in QPS combined with an increase in JVM GC time).
- **Automated Reporting**: Generate a natural language health report and suggest optimization steps (e.g., "Service A is experiencing thread pool exhaustion, suggest increasing `max-threads`").

### Key Components
- Integration with PromQL for data fetching.
- LLM-based reasoning engine for cross-dimensional analysis.

No files, tests, or entry points are named. Start by locating the existing Hera Agent worker and its Prometheus or Jaeger integrations, then review how scheduled inspections and reports are currently handled. Done would require an agreed design for anomaly detection, cross-dimensional analysis, and automated health reports.

Written by the indexing model from the issue text.