Database Health Check Monitor Designer

Effective monitoring is the nervous system of any high-availability database deployment. Without correctly designed health checks and alert thresholds, automatic failover systems trigger on false positives, real failures go undetected until users notice, and on-call engineers are buried in noise. This AI assistant helps database and platform teams design monitoring strategies that are accurate, actionable, and tightly integrated with their HA infrastructure.

The assistant produces health check specifications for every critical dimension of HA database operation: primary node liveness, replication connectivity, replica lag, replication thread state, disk space on data and log volumes, connection pool saturation, lock wait accumulation, and backup recency. For each check, it defines the metric source, measurement query or command, warning threshold, critical threshold, alert routing, and the recommended automated or manual response.

It generates monitoring configuration for common observability stacks — Prometheus with postgres_exporter or mysqld_exporter, Grafana dashboards, Zabbix templates, Datadog monitors, and cloud-native monitoring in CloudWatch, Azure Monitor, and Cloud Monitoring — producing configuration snippets, PromQL alert rules, and dashboard JSON where appropriate.

The assistant also addresses the critical relationship between monitoring and failover triggers. It helps teams define the health check parameters that their automatic failover system uses for primary detection, ensuring that monitoring alerts and failover thresholds are consistent rather than creating contradictory signals.

This tool is valuable for DBAs establishing monitoring from scratch, teams auditing existing alert configurations that generate too many false positives, and engineers designing the observability layer for a new HA cluster deployment.