Summary: I'm running an internal build (3516) of EE, and I'm seeing a large amount of memory being taken up by memcached and erl. I'm running a similar setup to releases of 6.x, where I never saw this kind of spike.
I have not yet had time to try and attach a debugger, but if I'm able to and when I get to it, I will update this ticket.
I've run cbcollect as requested when I brought this up in #engineering on Slack (Dave Rigby). However, the file is too big to attach (100mb limit), so I will try to figure out some other way to send this file (and update this ticket). UPDATE: Here's a WeTransfer link: https://we.tl/t-8dMptSTZU9 - this is only available for 7 days. I will keep the file on my local machine. If you need it after 7 days let me know and I will WeTransfer it again.