I don't think it's a bad thing to adjust. It's the DM's job to make the game fun for the group, which typically requires making it challenging.
I write custom monsters on a regular basis so I have no problem making changes to existing ones.
The difference is whether you change them
while they are in play, or
before they are in play. (Changing them after they have
left play is...not particularly useful, so I don't imagine you do much of that.) Making changes to something before it enters the play-space is perfectly acceptable. Necessary, even. Making such changes after it enters play, however, is neither necessary nor (I argue) acceptable.
If the players do their research and find out that the slime monsters are averse to cold, it is foul play on the DM's part to change their mind midway through the adventure and decide the slimes are
actually weak to acid, there just happened to always be acid involved in the places where the slimes were avoiding cold
without establishing why. If they do their research and know that owlbears never form groups larger than mated pairs and one or two juvenile offspring, then it is foul play on the DM's part to spring a pack of five adult owlbears on the party
without establishing why. Etc.
There is no meaningful difference between this and changing a monster's AC or HP values after those have already become tested by the players' actions (attacks made and damage dealt, respectively.) The players have acquired real information about those values, even if the precise number remains uncertain. To change those values without
potentially observable justification is foul play on the DM's part, even if the DM only does so for the absolute best of reasons. Cooking the books is still cooking the books if you do it to donate the gains to charity or (to use a real-world example) to help a sovereign nation try to work its way out of debt.
Note the word "potentially." The players do not have to succeed. But they must be furnished with the opportunity, and the opportunity must be genuine, not structured in such a way that it
technically is an opportunity but
practically ensures they'll never succeed. Players should be responsible for their own unwise choices, lapses of attention, or (some of the time!) sheer dumb luck because that's how flat-dice rolls work. But the chance should be there, and should be real, no "roll 3 crits in a row" BS to skirt the line.
---
Above, I said changing things before they enter play is potentially even necessary, and I stand by that. If the players
don't have any reasonable way of knowing how much HP a creature has, then there is no difference if it changes. Once they do, even if it is at a distance removed, they need to be able to know that their previous information is actually faulty, because
it wasn't faulty before but is now. For those aforementioned slimes, some kind of hint (or even direct statement from an NPC or evidence they encounter) that cold doesn't do anything or that acid is effective is plenty. For the owlbears, finding out there's an evil druid warping the behavior of animals in the forest, or that it's Owlbear Mating Season so groups of male owlbears sparring to attract mates, would be perfect evidence for contradicting a pattern that is
usually true
most of the time but happens to not be true in this specific case. Etc.
Covering these things up with mere fudging is doing both the game and oneself a disservice.