system-vigil_skill

Overview

This skill monitors host system health by collecting disk, memory, and CPU metrics and returning a structured JSON status for predictive maintenance. It highlights risk levels using flags like "warning" and "critical" so agents can act before resource exhaustion causes failures. The output is compact and machine-readable for automated workflows.

How this skill works

The skill runs lightweight Python code that invokes standard Linux utilities (df, free, uptime) to gather current resource usage. It calculates percentages, free gigabytes, and short-term CPU load, then evaluates those values against configurable thresholds to assign status flags. Finally it emits a JSON object with top-level status and detailed sections for disk, memory, and CPU.

When to use it

• Integrate into automation pipelines to gate deployments when host resources are constrained.
• Schedule regular checks for predictive maintenance and capacity planning.
• Trigger alerts or autoscaling when warning or critical flags appear.
• Collect historical snapshots for trend analysis and forecasting.
• Validate environment health before running resource-intensive jobs.

Best practices

• Run the check at regular intervals (cron or scheduler) to maintain a useful time series.
• Adjust warning/critical thresholds to match application sensitivity and host sizing.
• Combine outputs with log and metric stores for correlation and root-cause analysis.
• Use the JSON output as input to alerting systems or orchestration logic.
• Ensure the agent has permission to run df/free/uptime on the host.

Example use cases

• Pre-deployment gate: refuse deploys if memory used_percent exceeds a threshold.
• Auto-remediation: spin up a new instance when disk free_gb falls below critical level.
• Dashboarding: feed periodic JSON outputs into a monitoring dashboard for visibility.
• Batch job scheduler: delay heavy jobs when 15-minute CPU load indicates high utilization.
• Capacity planning: aggregate daily reports to project when more resources will be required.