How do you use LVM partitioning for good old VMs?

Question

linux

Good day.
During the development and widespread use of amazon web services, google cloud, and other cloud services, there are still old projects that are unlikely to be rewritten to work in the clouds, and are still running on the good old vps with a hypervisor that need to update the guest system.
1. Split LVM into subsections, for example: /home, /home/project, /var/log, /var/lib/mysql, /mnt/backup,
2. Keep minimal partitioning, for example: /, /mnt/backup ( in db, logs, etc., everything is together)
3. Make one section and do not take a steam bath.
Each of the numbers corresponds to a different approach, it is clear that everyone is different and each solution has its pros and cons. There is simply no one to ask how they prefer to do now. The first point saved several times when the section with logs overflowed and did not keep track of the monitoring, but the machine did not fall because everything else was partitioned or when it was necessary to expand the section with backups or data without reinstallation and migrations (projects then grow). The second way is something in between. Well, the third one is sometimes not convenient if you don’t need to extinguish the system, but you need to do some operations with disks or partitions (expand vg and add a disk, for example) when everything is not fault-tolerant.
Now I use all three approaches for different tasks, I think. But I want to do something more unified once the OS is being updated and the transfers are done somehow. With new ones, most often there is one lvm section and docker in it, or if the project is without it, then the third approach.

Answer 1

For the web
/
/home - applications here
/var/static - static that users load. noexec
/var - subds store their stuff here
/var/log - here if someone starts to shit in the logs, you will immediately notice, and applications will not fall due to lack of space
/boot
wo swap