robovacreferenceFive tables
Each one is a real report built from a real snapshot payload: the same route, the same formulas, the same sliders you would get from your own table. They are here because vacuum problems come in a small number of shapes, and reading five of them is faster than reading the documentation.
Two of the five cannot be fixed by any setting on the page. That is deliberate.
The default 0.2 scale factor on a 412 M-row table. Nobody set it wrong; nobody set it at all.
A table where the vacuum settings are not the problem. The fix is fillfactor.
Dead tuples are irrelevant here. The whole report is the freeze horizon.
The state no setting fixes. Every slider here is already correct.
A tuned table. The report says so and proposes nothing, which is the result most tools refuse to give.