Collect all cgroup metrics, tested only on CentOS.
metrics-cgroup.rb -g "/sys/fs/cgroup/**/**"
Collect only systemd service metrics
metrics-cgroup.rb -g "/sys/fs/cgroup/**/system.slice/*"
Collect only user.slice only
metrics-cgroup.rb -g "/sys/fs/cgroup/**/user.slice"
Collect only mesos task metrics
metrics-cgroup.rb -g "/sys/fs/cgroup/**/mesos/*"
Collect only docker container metrics (may not be meaningful if you already collect metrics from docker)
metrics-cgroup.rb -g "/sys/fs/cgroup/**/docker/*"
Collect only system.slice, user.slice, docker and mesos and nothing else
metrics-cgroup.rb -g "/sys/fs/cgroup/**/{system.slice,user.slice,docker,mesos}{,/*}"
Instead of collecting breakdown of each docker instance, or mesos task, or systemd service metric; this could let to collect/compare overall docker, mesos, systemd and user recources in total.
metrics-cgroup.rb -g "/sys/fs/cgroup/**/{system.slice,user.slice,docker,mesos}"