The model behind this report is a star schema. Sales is the fact table: one row per order, with numbers to add up and key columns such as StoreKey. Date, Product, Store and Customer are dimension tables: one row per item, with the text you slice and group by.
Relationships connect them. Each one links a key in a dimension (the one side, marked 1) to the matching column in Sales (the many side, marked *). Filters flow from the one side to the many side, which the arrow on the line shows. That's how Region on a chart's axis filters the sales rows.
When a relationship is missing, a visual shows the same grand total for every region, because nothing tells Power BI which sales belong to which region. Open Model view and drag the key column from Sales onto the matching column of the dimension. Keep the cross-filter direction Single unless you have a specific reason. Both also lets Sales filter the dimension, which changes customer and product counts in ways readers rarely expect.