Great question.
Here's my $0.05 (I'm giving you more than a tupence so deal with it).
If the question is old and good, then answer it if you can.
If the question is old and poor but can be edited to be good, then edit it and answer it.
If the question is old and off topic then vote for it to b...
@JohnK.N. Commented that follow upc is (part of accepted answer) 8s difficult since the "Last seejn" widget was removed! Any sign of that being reversed?
However, the last point in the list is pretty on point: Try to follow back up on it if you can for a while and if after a few months nobody has responded with more details, then vote for it to be closed.
As for the revision of the last seen option: I don't think so.
@ypercubeᵀᴹ - working on that as we speak! I know the fiddle is blank at the moment - I'm going to put in yours and also EB's without the RCTE's! Even playing field and all that! :-)
Would be nice to have a process that "vacuumed" all the old, unanswered questions - like, bump them once or twice, delete after a week if no responses.
I think for magic unicorn/useable answer perspective there isn't much motivation to answer a narrow question for a user who won't upvote/mark accepted and probably found a solution someplace else/just kludged it.
@McNets But I don't want to expend effort, I want things automated!
Even if the level of effort required to automate was greater than the effort to do things by hand because why else would I have gotten involved with computers? :)
@ypercubeᵀᴹ - revised benchmarks (numbers table permitted). 2 of your answers in - the one which was marked as correct and also the one which you put in the fiddle above yesterday. For some reason, dbfiddle just won't work for me with the 340/570 parameters. In a way, it's just as well - I have control over what's happening on my own machine (nothing for the most part). Anyway, my revised answer is here - with corrections - I'm hoping that it will only require very minor changes from now on!
Benchmarking - it looks easy, it ain't! I really admire that guy on Phoronix!
@ypercubeᵀᴹ - I've added a (absolutely final method) which uses the INTARRAY extension - this bit SELECT ARRAY_CAT(c.arr, SORT(c.arr, 'DESC')) - the SORT means that I don't have to do a second GENERATE_SERIES() and that seems to shave a few vital bits of a millisecond off the time! I think I've gone overboard on this now :-), so I'll stop there! Anyone would think that I didn't have a life! :-) Be very interested to see any improvements though...
Scratched the above because the OP initially said they were using SQL 5.7 - I assumed MySQL - apparently now it's PostgreSQL - so the provided answer should apply!