Of course, refereeing journal articles is an expected duty; I’ve published a few and therefore benefited from the service of referees.
And it is very important that referees do their jobs responsibly.
I’ve refereed a few articles and some were very easy to reject: they either contained gross errors OR contained proofs of items that were already well known…and the existing “known” proofs were simpler (e. g. appeared in widely read textbooks).
But the most difficult articles to referee are those that are both
1. Poorly written and
2. contain some content that might have mathematical value.
These sorts of articles are time-sinks; one has to read them carefully because those ideas might well be worth seeing in print…but my goodness they are painful to read.