with modern designs moving more and more towards very advanced SMPS, and ditching the traditional LDO and linear regulators, it's very hard to gain insight from an analog scope these days. Scopes have sort of transcended their traditional uses and limitations.
What did you have in mind? Analog oscilloscopes work fine for measuring various switching and recovery waveforms, checking transient response, and measuring RMS and peak to peak noise. DSOs can do these things as well except maybe for RMS noise. DSOs are exceedingly useful for measuring startup and shutdown behavior where a pretrigger record is needed; this can be done on an analog oscilloscope but it is much easier on a DSO and is one of the more common things I use DSOs for. DSOs are also better when waveform math is needed to measure quantities like power although I have done this with difficulty on analog oscilloscopes.
I don't think there is any real compromise in quality in Rigol PSU's. They are as good as anything out there.
I doubt they compromise either although they did have a recent problem with their DP832 power supplies. If their power supplies are just as good as everybody else's though, this could be damning with faint praise if everything else out there is designed as I described.
On the front of reliability, an older scope is going to be impossible to repair (in some situations). The only reason I can see for using anything (other than a good modern DSO/MSO) is for characterizing noise in highly critical power supplies.
I would consider any modern DSO to be an Apple like product after the warranty expires. If it breaks in a significant way except maybe for dried out aluminum electrolytic capacitors, then throw it away and buy a new one.
At least the older oscilloscopes have detailed service documentation which gives you a fighting chance at repairing them. Of course that would not help at all if one has to deal with complex parts using surface mount construction. On old oscilloscopes it is very difficult to repair hybrids although some people do it.
In a production environment, I just don't see how it's possible to spend the time to maintain old dinosaurs.
This is usually going to be the case with specific exceptions like when new instruments are not available which have old capabilities. If this happens then at some point the production process needs to be changed to accommodate new test equipment.
This is probably how I got my sampling oscilloscope in such good condition. Apparently it sat in a closet for 10+ years because nobody could figure out its simple failure or wanted to pay to have it repaired and was considered too expensive to throw away even if it was useless.