My first (of many) AI agentic workflow that automatically learns from my homelab hosted services on Proxmox and Portainer (+ other docker hosts) and checks for missing updates, misconfigurations, resource stagnation, and many other metrics which then correlates with logs and extra metrics from Zabbix, Grafana and Prometheus, and neatly infers what steps should be taken if any for each host, service and issue. I set it up to send me notifications through NTFY but also creates tasks with summary of issue and proper cataloguing.