This could be a sore subject with many a end-user. System Administration takes time, a desire to keep up, and the knowledge to maintain one's system.
1. Many users just want to use the computer. That attitude reminds me of a person who drives a car until it stops and then buys another. It is costly, wasteful, and counter-productive for end-users, businesses, and yes, even the industry. If you're replacing hardware because you didn't maintain the Operating System crashed and you didn't bother to learn it well enough to take care of the issue. Shame on you. 
2. Different hardware has different strength and weaknesses. Reading about what those areas are and how they effect performance, allows the Administrator or end user to head off an issue before it kills a perfectly good OS. Example: Intel 32 Bit processors tend to bugger the release of memory at the end of application execution. This causes memory fragmentation. Solution: (For Linux) Create ".so" which acts as a traffic cop, releasing the fragments (memory address) from the stack. Options to do so exist as far back as the memory management calls in dBase II (old RDMS). Linux offers easier solutions as does Windows, it's called reboot the system at least once a day (end user), once a shift for businesses, or when you notice the system slowing when you change from one application to another. Failure to learn about why you system isn't responding normally is pure laziness.
3. fsck or scandisk and defrag.... Today's systems are solid enough that these are rarely used by an end-user. However, disk fragments can be released via their use. Using cron to schedule their use in the empty hours of the early morning once a week could solve the issue. Puppy boots as root, which makes it simpler, otherwise scheduling is automated in Slackware, Arch, Suse, Ubuntu (and its flavors), Debian, and most distributions. It means that an UN-attended reboot should also be scheduled.
4. Home setting: Many users turn off their computers, advise then to get a UPS system or at least a surge-protector, and leave the system up, at least one day a week. Maintenance is cheap! New hardware is costly. Use cron or the Windows scheduled to set up tasks like clearing temporary files, thumbnails, browser cache, system dumps, and/or do backups.
5. Cloud storage is less than secure. Use local usb external drives,thumb drives, CD/DVDRW drives as the asset they are. Storing a daily backup in a drawer is simple and effective, offsite weekly backups kept somewhere other than one's home is as simple as asking family or friends to keep an encrypted backup copy, is easy and usually painless. Loosing one's data is painful. House Fires, Tornado's, Floods, power surge damage, or even the family dog knocking over a desktop pc happen. Protecting one's data is just smart. I keep an old Pentium II system on my home network with shared raid removables simply for my systems, my wife's systems, my sister-in-law's system (lives across the street), and my daughter's system (Lives 2 house north) to catch daily scheduled backups. Swap the re-movable's daily and it's done. (Again keep the systems up while you sleep and let them do it for you)...
6. Finally, Databases (small or large) are money. maintaining them requires little time using Cron, AT, or Windows scheduler. Most usable RDBMS systems (Less MS Access) have built in utilities to clean up the deleted, batch db defrags (contiguous databases run faster), and index rebuilds can all be scheduled. Do it! Recover time is wasted time and money.
Enough for this Rant. My companies software support base has grown 300% even during the pandemic. Most or all the costly issues, which my clients endure, could have been avoided by doing the above Administration.