You're correct in that you can only test for things you know about (duh, you don't know what you don't know) so as far as screening goes, it's just a matter of how much you're screening FOR. It's also really easy to rule out which diseases are not likely to occur just by checking one person's DNA. You only have to screen where both samples can potentially combine to form something bad.

Far from all hereditary disease are easy to rule based on genetic tests.

You need to differentiate between Mendelian and Complex genetic disorders. While genetic tests are useful to indentiy disease caused by a single allele (Mendelian disorders), they today are inconclusive for many other Complex diseases, like rheumatoid arthritis, with a known genetic component. These disease are multifactorial: the interplay between several genes (variants) and environmental stimuli cause the disease. One of several genetic mutation must be present for the disease to occur, but without an environmental stimulus, the disease will never materialize. As such, even though we have acces to a full genom sequence, genetic tests are inherently limited and thus inconclusive in determining risk for many diseases.