Build a useful investigation brief
Freeze a click-path fixture
Choose a parent state with a known dimension value and record filters, period boundaries, grouping, sort order, and expected child population. Save the exact click sequence and a small data fixture that contains both matching and nonmatching rows.
Diff the scope inputs
Compare parent and child query inputs line by line: predicates, joins, date ranges, selected columns, grouping keys, and limits. Explain each difference as inherited, intentionally narrowed, or unexplained. Note whether the child is a record list or another aggregate.
Write the causal boundary
State the earliest scope transition that changes the answer, then link it to the observed child result. Distinguish evidence from a hypothesis about implementation. Include a counterexample where a different clicked value would reveal whether the behavior is systematic.
What to carry forward
The investigation is sufficient when another reviewer can reproduce the click and identify the first changed scope input. Stop short of naming a root cause if the available evidence only shows a contract mismatch; preserve the exact parent state and child comparison for the owner.
Technical background: Metabase documentation.
Keep the decision with the work.
Use a Work Item in Aglet to record the problem, the evidence you have, and the next decision. Add an owner and priority, then keep updates in the discussion so the next person can follow the reasoning.
Create an account See the product workflow