I had similar issue with BTRFS thanks to new faulty DDR4.

- CRC checksum for some blocks on fs was wrong

- I discovered faulty DDR4, replaced it, without knowing FS corruption

- system started randomly kernel panicking

- no msg in logs (fs died with kernel)

- eventually I experienced kernel panic while at computer, and seen msg on screen (broken checksum or something)

- there was no way to repair filesystem. I had to recover as much as data, throw away some files and start over on fresh

- I use BTRFS snapshot streaming in backups, wrong checksums propagated to backups. I had to restore from 1 week old backups

- Since than I use RSYNC backups as well

- And I test all PCs with memtest monthly overnight. It is also a good stress test for CPU.