One advisor surface for mixed database estates — consistent findings, one runbook, one place to triage. Browse the rule catalog.
99 deterministic advisory rules
Performance, capacity, hygiene, and reliability checks across all five engines — translated into ranked findings with severity, confidence, and evidence attached. No AI in the critical path; findings are reproducible and auditable.
AI DBA agent — in your environment
A tool-calling agent investigates issues step by step: it analyses query patterns, recommends engine-specific index DDL, estimates impact from baselines, and drafts deployment runbooks. Powered by an in-environment model — no third-party LLM, no egress. Bring your own provider if you prefer.
MCP-first by design
Every capability ships a matching MCP tool — 33 and counting. Point Claude, Cursor, or any agentic client at InsightralTM and query findings, schemas, baselines, and runbooks directly from your editor.
Query intelligence with cost attribution
Top queries by total and mean time, normalized and redacted so literals never leave your database. Per-statement latency regressions, plan-change detection, ghost-query detection, and a per-query dollar cost model across CPU, IO, and temp usage.
Schema map and drift prediction
An interactive graph of tables, columns, and foreign keys with full change history. Predictive drift detection scores per-entity risk on a 30-day horizon — before a migration surprises you.
Incidents with root cause built in
Findings correlate into incidents that carry the whole story — database, finding, alert, probable cause, and the recommended fix — plus runbooks that can execute with pre/post metric comparison and automatic rollback.
Forecasting before the pager fires
Capacity prediction on up to a 360-day horizon, threshold-breach forecasts per database, and fleet-level anomaly correlation that surfaces risk before an incident reaches your on-call rotation.
Alerting you can route
Per-rule dispatch to Slack, Microsoft Teams, webhooks, and email. Your team decides who sees what, when.