Well I have been talking about this for some years now and looking at what others are doing. I'm ahead of the game here and have been for some years.
I noticed over 40 years ago that when a component failed, it was replaced and thrown in the bin. Electrolytic caps of course being the common problem. But who ever bothers to check why, where is the forensic examination of the component, who has the interest or time to do it ?
Then I got interested in why electrolytic caps were failing, some were young (<5yrs) others older >40 years.
There are a number of mechanisms of failure at least 4. However the common one is that the rubber seals allow water vapor to escape (even if they do not allow the frank electrolyte to leak as it often does). The capacitor goes high ESR and low capacity and loses physical mass. I have confirmed this many times, that the defective capacitor can also be returned to its original capacity and ESR value, by soaking it (out of its canister)in deionized water for 24 Hrs.
In any case, because of these issues, including the very destructive effects of leaked electrolyte, in my designs & restorations of tube gear now I seldom if ever use electrolytic capacitors in the high voltage power supply sections. Even the re-builds of vintage gear I go for polyester or polypropylene caps. One advantage is that vintage electrolytic caps were very physically large for the capacitance and voltage rating, and modern non electrolytics of the same uF and higher or equal voltage ratings, quite compact.
Have a look at the MKP capacitors I just used in this project, to avoid electrolytics:
viewtopic.php?f=12&t=371368And also the photo attached of a multi section electrolytic, from a vintage TV restoration, re-built with non electrolytic capacitors, which have the same uF voltage but actually higher voltage ratings than the original electrolytics.
This might not be an economic option, but good things rarely are.
It is quite common to do a restoration on tube apparatus and if its done with electrolytics, to see them fail again in less than 5 to 10 years. So it needs a different approach, unless you think that is ok, I don't. In the service industry it is thought to be ok, because that time frame exceeds any common guarantee period, but what if it is for your own equipment I wonder.