In the last two months, the weekly cost of smtp servers has almost doubled.
Within the billing history, I've found strange numbers
starting from the last week. See the attached screenshot.
The old weekly billing was approx. XXX per week,
starting from October 26 it rised to approx. XXX per week.
The systems usage by our side remained the same.
Could you please check it?
On November 5th, I sent the above message to my major hosting provider.
It may seem like a small problem, but the weekly cost has grown at unsustainable rates
and I had to keep paying it to avoid service interruptions.
After many messages and approx. two months, the issue was finally closed:
I was able to confirm on multiple containers that the sessions
written on the /run ramdisk are causing the memory leak.
I suggest you update systemd and dbus on all containers
and if looking to maintain a minimal ramdisk usage,
have a cron job clearing scope sessions every hour.
I had to perform a system update on all servers
and add two clearing commands to the daily cronjob activity.
On two servers, the ones with the greatest load, it wasn’t enough.
A daily reboot fixed the system overload.