Native parameter support at the dashboard level

What is the problem or goal you're trying to solve or accomplish?

Users need a way to drive calculations, metric toggles, and axis definitions through interactions, without filtering rows. There is currently no concept of a dashboard-level variable that feeds into chart logic independently of filters.

How are you solving it currently?

Jinja templating with filter_values() in virtual datasets can partially replicate the behavior, but requires virtual datasets and is not triggered by clicking a viz mark. Cross-filtering only propagates row-level filters, not logic changes.

What is your recommended solution?

Introduce a native parameter concept at the dashboard level, a typed variable (string, number, date) that can be set via filter panel or chart interaction, and referenced in chart configurations and SQL logic without filtering rows.