Regularly scheduled pull + reboot/re-deploy
Partly to get code into the live version and partly to hammer out all the pieces of a pull and re-deploy so we can have really solid process for that and predictable schedule for when code changes go live.