I want to use --cgroup options with perf to collect performance metrics within a container.

I created the cgroup in the following way:

cgcreate -g cpu:/perf_test
cgcreate -g perf_event:/perf_test
cgexec -g "cpu:/perf_test" -g "perf_event:/perf_test" stress-ng --cpu 16 --timeout 1y

I use perf in the following method, but it keeps failing:

# perf stat -e cycles,cpu-clock --cgroup=perf_test
Usage: perf stat [<options>] [<command>]

    -a, --all-cpus        system-wide collection from all CPUs
    -A, --no-aggr         disable CPU count aggregation
    -B, --big-num         print large numbers with thousands' separators
    -C, --cpu <cpu>       list of cpus to monitor in system-wide
    -c, --scale           scale/normalize counters
    -D, --delay <n>       ms to wait before starting measurement after program start
    -d, --detailed        detailed run - start a lot of events
    -e, --event <event>   event selector. use 'perf list' to list available events
    -G, --cgroup <name>   monitor event in cgroup name only
    -g, --group           put the counters into a counter group
    -I, --interval-print <n>
                        print counts at regular interval in ms (>= 10)
    -i, --no-inherit      child tasks do not inherit counters
    -n, --null            null run - dont start any counters

What could be the reason?

The perf version is: perf version 3.10.0-693.5.2.el7.x86_64.debug

OS version is: CentOS Linux release 7.4.1708 (Core)

kernel version is: 4.18.0