How do I detect anomalies in my metrics with iDBQuery?

Anomaly detection finds values that fall outside a metric's normal, expected range, an early warning that something has changed. Rather than staring at dashboards, iDBQuery lets you ask for the anomalies directly and monitor many metrics at once.

• You'd ask iDBQuery: 'Compare this week's daily orders, revenue, and refund rate to their trailing 8-week pattern, and flag any day that fell outside the normal range.'
• It measures each metric's typical range, compares recent points, and returns a chart with the outliers marked plus a table of what deviated and by how much.
• Every flagged point is cited to its source rows, so an anomaly is verifiable, not a mysterious alert.

This differs from investigating one known anomaly: here iDBQuery scans across metrics to catch changes you were not watching for. Once something is flagged, you can pivot straight into diagnosis in the same thread: 'Break the anomalous refund spike down by product and region.' Because the metrics may span several systems, the one live model lets it watch across all of them with no pipeline. Turn it into a scheduled report so anomalies are surfaced automatically, and let the Analyst agent investigate a flagged movement on its own, citing the rows that explain it. This is transparent monitoring: the SQL and the ranges used are always visible.