devops-monitor_skill

Overview

This skill provides a focused DevOps monitoring agent for containerized services, logs, healthchecks, and alerting. It centralizes commands and checks to quickly assess container state, service health, network exposure, and common error patterns. Use it to triage incidents, confirm rollbacks, and maintain uptime SLAs.

How this skill works

The agent inspects Docker containers, images, volumes, and networks using targeted CLI commands and extracts recent logs for analysis. It runs healthcheck requests against live and ready endpoints, queries systemd service status for key services, and scans ports and network resources. Alert levels are mapped to concrete responses: immediate rollback for critical failures, investigation for warnings, and routine monitoring for informational states.

When to use it

• Triage production incidents involving containers or APIs
• Validate deployment health after a release or rollback
• Investigate error and warning patterns in application logs
• Confirm network port bindings and Docker network configuration
• Automate basic monitoring checks in runbooks or CI hooks

Best practices

• Run container status and resource checks (docker ps, docker stats) before log analysis to scope issues
• Use health endpoints (/health/live and /health/ready) as the primary availability signal
• Filter logs for error and warn patterns before deep dives to reduce noise
• Correlate systemd service status with container state for service-level failures
• Apply alert thresholds: treat healthcheck failures as critical and response delays >2s as warnings

Example use cases

• Check running containers and resource usage for the nest-api service with docker ps and docker stats
• Fetch the last 100 application logs and grep for error/warn to identify recent failures
• Run curl health endpoints to decide whether to rollback or continue a deployment
• Inspect ports with ss and confirm Docker network presence when troubleshooting connectivity issues
• Query systemctl status for Caddy and Docker when external routing or container runtime issues surface