Never silently widen a dashboard's selected time window

When a metric has no evidence inside a user-selected time range, never backfill it from a broader range while leaving the selected-range label unchanged. Show the metric as unavailable and disclose the actual evidence coverage. A fixed capture subwindow may legitimately produce identical values across multiple larger filters; label that capture interval explicitly so the equality is explainable. A fallback is acceptable only when users explicitly opt into it and the substituted range is shown beside the value.

Applies when: Interactive analytics combine operational windows with partially captured communication or audit evidence. Keywords: dashboard date filter missing data silent fallback coverage fixed sample