I’m trying to debug a test suite where processes suddenly disappear. I would like to check if any process got killed because docker’s cgroup is out of memory. Previously, dmesg
could be used to to this, but this is not allowed, anymore:
circleci@13ab86076843:~$ dmesg
dmesg: read kernel buffer failed: Operation not permitted
Is there a way to call dmesg when sshing into a test machine? Is there another way to check if the OS killed any processes?