Higher idle process activity after upgrade

Such are the things I see in top sometimes:

    PID USER      PR  NI    VIRT    RES    SHR S  %CPU  %MEM     TIME+ COMMAND                                                                     
   1024 discour+  20   0  333584 202888  21660 S   1.3   2.5   3:35.81 ruby                                                                        
   1156 discour+  20   0  715904 249416  24532 S   1.3   3.1   3:26.50 ruby                                                                        
   1178 discour+  20   0  726664 251468  24564 S   1.3   3.1   3:26.11 ruby                                                                        
   1189 discour+  20   0  714368 247912  24444 S   1.3   3.1   3:24.56 ruby                                                                        
   1200 discour+  20   0  709760 249708  24632 S   1.3   3.1   3:22.96 ruby                                                                        
   1234 discour+  20   0  713344 250288  24636 S   1.3   3.1   3:30.24 ruby                                                                        
   1167 discour+  20   0  712832 247928  24436 S   1.0   3.1   3:24.75 ruby                                                                        
 188658 me        20   0   10424   4240   3576 R   0.7   0.1   0:00.36 top                                                                         
    448 root      20   0 1748444  84900  45884 S   0.3   1.1   6:09.98 dockerd                                                                     

with no one connected…

The 3.5 minutes accumulated CPU time each is after 35 hours up, with very little user activity.

I also tried to use rbtrace, but it complains:
(pid is not listening for messages, did you require “rbtrace”)
Do I need to rebuild the image to fix this? Or can I just adjust or rebuild something inside the container?