Host & System Management Skill

This skill provides direct visibility into the physical server's resource health and the status of systemd-managed hub components.

🛠 Available Tools

Tool	Purpose	Input Schema
hub_inspect_host	Inspect physical resource pressure (Load, Memory, Disk)	{}
hub_list_host_services	List and check status of core systemd units	{}
hub_query_service_logs	Query systemd journal logs for a service	{ "service": "string", "since": "string" }

📋 Standard Workflows

1. Resource Pressure Investigation

If services are slow but metrics look okay, run hub_inspect_host to check for physical bottlenecks:

1. CPU Load: Check load averages (1m, 5m, 15m) relative to core count.
2. Memory: Check available vs. used memory.
3. Disk: Ensure partitions (especially for databases and logs) are below 90% capacity.

2. Service Management

For core hub components (Ingestion, Proxy, OpenBao) that run as systemd units:

1. Run hub_list_host_services to ensure all units are in the active (running) substate.
2. Use hub_query_service_logs with since: "5m" to view recent logs if a service is restarting.

💡 Operational Tips

• Core Units: Common units include ingestion.service, proxy.service, and openbao.service.
• Time Window: Standard formats for since include 10s, 5m, or 1h.

---

For detailed API documentation, see references/api-specs.md.