I have been reading this old post[1] again, and am worried that developers are just doing something without thinking. (Don't whine about such statement, because that is quite common these days) I honestly do not get why my default setup is causing systemd-journald on a host to go 100% while I increase the logging of a container that logs remotely (via /dev/log)
The only way to disable the waste of resources is to disable logging. But afaik then issues with starting services on the host are not reported as well. What kind of weird design is this? This makes me start to doubt the quality of logics applied in this design. Does anyone in your team even test on things like changes that cause twice the amount of iops? [1] https://www.mail-archive.com/systemd-devel@lists.freedesktop.org/msg47844.html
