You've built your AI, tested it thoroughly, and launched it to users. Now comes the most important part: making sure it actually stays good. The Production Dashboard is your command center for monitoring live performance, catching quality issues before they become problems.

Accessing the Dashboard

Find "Dashboard" at the top of your sidebar under "OVERVIEW." This is often your landing page when logging in.

The dashboard displays two main tabs: Performance Metrics (default) and Summary Dashboard. Most monitoring happens in Performance Metrics.

Top right controls include a date range selector (defaulting to "Last 30 days"), a Filters button, and an Export button for downloading reports.

Reading the Performance Trend Chart

The main visualization shows evaluation activity over time with three colored bars for each date:

Blue bars represent Total evaluations run that day. Height indicates volume.

Green bars show Passed evaluations that met your quality thresholds.

Red bars indicate Failed evaluations that didn't meet thresholds. Small red bars are good. Tall red bars mean quality issues.

The ideal pattern is tall blue bars with equally tall green bars and tiny (or no) red bars. This means active testing with everything passing.

Customizing the View

All Series dropdown lets you filter to specific validators. Focus on Toxicity trends, Hallucination patterns, or other specific metrics.

Daily dropdown changes time granularity between Daily, Weekly, or Monthly views. Weekly and monthly views help identify longer-term trends.

Date range selector ("Last 30 days") lets you switch to Last 7 days, Last 90 days, or custom date ranges.

The testing and security work you did earlier ensures your AI launched well. The Production Dashboard ensures it stays that way. Keep watching these metrics and you'll maintain the quality standards you worked hard to establish.