Once a day or two we get the bellow error and the mysql crashingt:

[ERROR] [MY-012872] [InnoDB] [FATAL] Semaphore wait has lasted > 600 seconds. We intentionally crash the server because it appears to be hung.

And before are alot of those:

Mutex at 0x7fde903975e8, Mutex TRX_SYS created trx0sys.cc:565, locked by 140504716736256
--Thread 140504724723456 has waited at trx0sys.h line 598 for 240 seconds the semaphore:
2024-07-13T04:43:26.039036Z 0 [Warning] [MY-012985] [InnoDB] A long semaphore wait:
{}

I also see a lot of those warnings in the logs:

[Warning] [MY-013865] [InnoDB] Redo log writer is waiting for a new redo log file. Consider increasing innodb_redo_log_capacity.

I couldn’t figure out what is the cause for this problem.
I increased the innodb_redo_log_capacity by a lot but it happend anyway. I also tried turning off innodb_adaptive_hash_index but it didnt help.
I thought of increasing innodb_log_buffer_size which is currently set to 1M because I thouth maybe its causing a transaction to be to slow.