Unified health monitoring dashboard for the Pelagic Fleet.

• lighthouse.py — FleetLighthouse class with health checking, status transitions, dashboard rendering, and diagnostics
• health_store.py — Persistent JSONL health data store with query, statistics, export, and pruning
• alerting.py — Alert management system with configurable rules, deduplication, and multiple channels
• cli.py — Command-line interface with serve, status, history, alerts, export, and doctor subcommands
• tests/test_lighthouse.py — Comprehensive test suite (48 tests)

# Start live dashboard
python cli.py serve
# One-shot status check
python cli.py status
# View health history
python cli.py history "Keeper Agent"
# View alerts
python cli.py alerts --all
# Export data
python cli.py export --format json
python cli.py export --component "Trail Agent" --format csv --output trail.csv
# Run diagnostics
python cli.py doctor

• OK → 3 consecutive failures → DEGRADED
• DEGRADED → 10 total consecutive failures → DOWN
• DOWN → 3 consecutive successes → DEGRADED
• DEGRADED → 3 consecutive successes → OK
• UNKNOWN → first success → OK