After the last tenant sync we notice the chrome frontend pods were running out of memory and restarting rapidly. To stop the bleeding we've added a job to cycle the pods every now and then, but we should dig into the root cause of the issue. This is not happening in commercial.