A post that violates their terms of service, which is a legally-binding contract that you and every other user on the platform agreed to when you signed up. Businesses have various motives for making users sign these agreements. The obvious purpos is to cut down on frivolous litigation, but they have other reasons, such as protecting the reputability and consumer favorability of its products.
Some products like Cabela's or Remington firearms probably couldn't give two shits if liberals get outraged over something they say or do that's offensive, but FB is a different beast in that it is trying to appeal to a broad cross-section of the population. So it has to walk a fine line between allowing a lot of the far-right rhetoric that conservatives use to trigger and troll the libs while simultaneously not letting it go too far (i.e, potentially crossing over into hate speech outright).
As I said in another post, I think the problem that Facebook, Twitter, and others created for themselves is that they were not consistently enforcing their TOS, so it's actually understandable (to a certain extent) that conservatives would claim they or those they follow were canceled. In the case of Trump, for instance, he had said all kinds of things over the years that were borderline judgment calls. And to be completely objective, the very post(s) that got him sent into Twitter exile were relatively tame. It's pretty clear that it was the timing of Trump's tweets - in conjunction with their tone-deafness - that ultimately got him shit-canned from the platform.
I would be willing to have a truly bipartisan review of social media legislative/regulatory policy such that it puts greater emphasis on social media companies demonstrating consistency in applying their standards. These companies collect data - that's their core business, so there's surely a way that can compile data that shows the degree to which they are applying their own standards. They just have to have the desire to do it.