Compare COUNT(*) to COUNT(DISTINCT pk), or row counts before/after the join — a jump means multiplication.

Before trusting any joined aggregate, interrogate the grain: count rows, count distinct keys of the side you're about to sum. Equal → the join preserved grain; unequal → the many side multiplied you, and every parent-level SUM downstream is inflated by exactly that ratio. The same probe works preventively (before writing the aggregate) and forensically (the revenue looks high — check the counts first).

SELECT
  SUM(g.kilos) AS kilos
FROM
  grow g
  JOIN harvest h ON h.batch = g.batch;

SELECT
  SUM(kilos) AS kilos
FROM
  grow;

The joined SUM read 40 kilos, not 30 — batch B1 had two harvest rows, so the join multiplied it; COUNT(*) was 3 where COUNT(DISTINCT batch) was 2, and that gap is the inflation.

The two-count probe, plus GROUP BY pk HAVING COUNT(*) > 1 to see the multipliers.