The grouping keys define what one output row represents. Grouping a repair log by Workshop gives one row per workshop. Grouping by Workshop and Shift gives one row per workshop-shift pair. The choice changes the question the result can answer; it isn't just a display preference.
Choose Advanced in Group By to add grouping columns and more than one aggregation. Name each result for what it measures. Row count counts log entries, while a distinct count of Ticket counts separate tickets. Several time entries can belong to one ticket, so those values need not agree.
Nulls deserve separate attention. A missing Workshop forms its own group. A missing Minutes value contributes no known time to a sum and must not become an invented estimate. Keep those records visible when the request covers the full log. Select the keys and measures from the business question, then check that the output grain matches it.
Table.Group(Source, {"Workshop", "Shift"}, {{"Logged Minutes", each List.Sum([Minutes]), type number}, {"Entries", each Table.RowCount(_), Int64.Type}})
Return one workshop-shift row with two measures that describe different things.