• You see high CPU utilization on worker and primary nodes when running bosh vms -–vital.
• vRealize Log Insight is configured in PKS tile without any rate limit under the logging section.

• You see output similar to the following when running a command similar to bosh -d <service-instance_xxxx> ssh <worker/xxxx>
  <worker/xxxx>#top
  
  Tasks: 152 total, 1 running, 118 sleeping, 0 stopped, 0 zombie
  %Cpu(s): 91.4 us, 8.3 sy, 0.0 ni, 0.0 id, 0.0 wa, 0.0 hi, 0.3 si, 0.0 st
  KiB Mem: 2041816 total, 251604 free, 1444388 used, 345824 buff/cache
  KiB Swap: 2040828 total, 1324040 free, 716788 used. 348904 avail Mem
   
    PID USER      PR  NI    VIRT    RES    SHR S %CPU %MEM     TIME+ COMMAND
  10529 root      10 -10 1692720 826192   5600 S 90.4 40.5  24927:17 ruby
   9899 root      10 -10  829624  63144  12636 S  3.7  3.1   1608:37 kubelet
   9322 root      10 -10  378708  30128   1376 S  1.3  1.5 645:01.23 dockerd

• You see output similar to the following when running a command similar to <worker/xxxx>#ps ax | grep ruby
  
  10543 ?        S<l  17553:36 ruby /var/vcap/packages/fluentd/bin/fluentd -c /var/vcap/jobs/fluentd/config/fluent.conf -p /var/vcap/packages/fluentd/plugins/ --no-supervisor
  28356 pts/0    S+     0:00 grep --color=auto ruby

This issue is observed when Worker VM Type or Primary VM Type selected as small(cpu: 1, ram: 2 GB, disk: 8 GB) in plan in PKS Tile and vRealize Log Insight is configured in PKS tile without any rate limit under the logging section.

​​​​​This is a known issue affecting Enterprise PKS.

One of the following options can be employed to workaround this issue:​

• Apply a rate limit to the vRLI settings in PKS tile under the logging section. See the instructions at Installing PKS on vSphere with NSX-T for more information. Apply the changes and validate that the CPU utilization is lowered.

• Increase the CPU count of the worker/primary node in the plan by changing the VM Type in the PKS Tile.