I recently moved from Florida to Indiana (I know… WHY???). I’m not used to these decorative-looking iron porch “supports”, but they seem to be everywhere on the older houses. Home Depot carries one with “0” rated load support; Lowe’s offers some from Gilpin that are load rated. So, actually 2 questions: 1) Is there any way that a porch roof could be safely cantilevered without support, BEFORE the advent of engineered trusses? And 2, assuming a given porch should be supported, how can I tell if the irons used are structural grade, or merely ornamental; anything better than tapping on it and guessing? Note: photos are from listings, not inspections, so I don’t have any additional info on these. And yes, obviously the one was done poorly; it’s sitting on the brickmould. LOL
Neither of these would trigger a comment from me unless the posts are damaged or bowing.
Yes of course.
By its performance perhaps. 
Brian, that’s kind of where I was leaning- so don’t call out the supports, or lack of, but look for other signs of failure instead. Thanks!
Yes, but not on the structure in your pic.
As far as the post, I would look for corrosion at the base of the posts.
I see cantilevered porch covers all the time. Basically roof trusses that extend over the porch. Kinda funny that they bother to put columns that appear to be supporting the cover, but the porches have settled, and the columns are just standing there with a gap usually at the top (nails keep them from falling), but sometimes they are hanging from the cover.
I’ve always assumed (maybe incorrectly?) those tubular metal things were load carrying. I definitely look for damage… which is common. It’s almost impossible that over the last 70 years some drunk guy with a lawnmower hasn’t rammed the thing. With no damage and no performance problem (sagging roof, etc.) I don’t comment.
Agree with the others, but that second picture would have me questioning. It looks like there is a sag in the middle (left to right), and those ones are not cantilevered. They are attached to the wall. So the other end would have to have a header across from post to post, but that is highly unlikely, given the thickness, and then the fact that it extends past where the header would be.