Improve documentation for downtime free updates
Right now our documentation for performing an update without downtime is fairly confusing. It starts off discussing post-deployment migrations without first describing what this is and why it is important. Next it moves on to an example with Chef as a deployment method, but this is going to be an extremely small subset of our customers. Most will use the packaged Omnibus scripts.
We should improve this documentation to tailor it to what the customer wants to achieve, upgrades without downtime. So the title of the page should describe that feature, then it should describe how it is achieved (post-deployment migrations) and then we should have an example and instructions for performing this with Omnibus.
Activity
WDYT @marin?
We already had a MR to improve the situation a bit https://gitlab.com/gitlab-org/omnibus-gitlab/merge_requests/1646 . We can build on top of it if needed.
