What you mention doesn't sound like container problems. To be frank, dropping hotfixes and ssh-ing into production machines sounds like you have a rather un-systematic approach to ops and containers now show it to you.