Like many things in ttrpgs, I think the metric for "is x good" or "should I x" is does it create fun or interesting choices. A tell is one way of presenting an interesting choice with a trap: there might be some player skill in recognizing the tell (so, a choice to interact), choices on whether and how to try to disable etc. Of course, in isolation, the most basic "obvious tell -> roll to disable" setup is maybe not the most interesting without other factors, like real risk, time pressure etc. But the general situation of type of "here's this obvious trap, got to figure out how to disable, and if it is worth it to interact at all" is great.
There are others that work well with no tell though, like "this is a thing that somebody likely would trap" (like a wizard's chest or a defended chokepoint), or a tell-less trap puts players in an interesting situation (like a portcullis splitting the party). The biggest failure mode of tell-less traps being common is overly cautious play (where in the worst case, there are not really choices, just the party declaring that we check this room for traps for every room). But this can be helped if caution is a scare recourse to be applied judiciously, usually by time being a factor, meaning there are big costs to cautiously checking everything, so you have to be smart and choose the most likely things to be trapped.
