These links won't be available for people outside PIH Jira

yeah these links were more of a placeholder to remind me to pull out information from the ticket. but i'm just going to comment them out for now.

what about any data that was inserted since the backup was taken? Not much point to making it rolling deployable if everything done during the deployment is lost. Maybe I'm misunderstanding though

we talked about this over slack. I was assuming this "parallel" option meant rolling deploy / no downtime deploy. Really it just means installing a new instance in parallel then migrating over when you're ready.

I think what's throwing things off for me is the "Switch Over" step is glossing over quite a lot. I think what you're suggesting is taking two backups: One at the start of the process that we can upload to the new 0.9.x instance whenever we're ready to test the instance, and then a second backup while actually switching over to the new instance while the old instance is down to capture any new data that was added since then.

So we'd want the "Disable Production Environment" step to be before we do the second backup (not sure if you're still writing all this out).

to the point about "parallel" vs "in-place" migration, i am using those terms because those were the most common. however i still don't like "parallel" as it causes confusion. my first attempt at this was "clean slate" vs "in place", which i don't like because "clean slate" carries a conotation that is not true (i.e. we're starting fresh, including the database). but now i'm also seeing that "side-by-side" is also used to describe "parallel" migrations, so i might use that breakdown instead.

regarding the "switch over" part, i'll just say that in general there's a lot missing so i might need some help structuring the missing buckets and maybe even writing some of these sections because my brain is mush after working on this for so long.