I *think* this is how I avoid "oomkill fights"
But I am wondering if there is any better way to get and view coherent logs/reports about oomkills. I think I could make something like this by writing a python script that scans dmesg at all times and gives me a daily report like "mysql oomkilled 8 times today, apache2 once". It seems like something that might exist already (but if it's vibecoded, I'd rather write the python script myself). (3/3)
@mcc /proc/vmstat has a line `oom_kills` counting the number of OOM kills since boot.

