Wiki and Issue Tracker migration and support impact

[Cantor, Scott]

Hello,

I sent a more detailed version of this to the Consortium Members list today, this is a stripped down version. I will be posting some version of this as well.

With Atlassian sunsetting Confluence and Jira Server in 2024 and given various trade-offs in the timing and other considerations, we're planning to perform migrations of our systems to the Cloud versions during the rest of 2021.

This is going to be disruptive, there's no way around that. Among the most visible issues:

- We lose practical support for doing federated login to either of them.

- Related, the Cloud versions are purely driven by email-based identifiers and the conversion will create assumed Atlassian IDs that match emails in the profiles now. Which accounts people actually use if they need access in the future don't have to be the same as the migrated ones of course.

- Unfortunately, the URL pattern for the wiki in the Cloud version is not the same as the format used for the Server version, so auto-rewrites of the requests to the old site simply won't work. The best I have come up with is to redirect the original requests into a transition page and then on to a wiki search for that topic in the Cloud instance. Jira issue URLs do appear to allow us to do an automatic rewrite when that time comes.

- I do expect that non-read access to the sites is going to be less open simply because provisioning is much different in the Cloud without the SSO -> auto user creation model we have now. We are likely to grant anybody with a non-generic email (i.e. not gmail) access but doing that manually is ugly so I really just don't know what we will do yet. Access to Confluence is very likely to be more limited simply due to less need, but those that have helped fix documentation will certainly be welcomed to do so. We may end up with some kind of invite-based model so those with access can invite others.

What do you need to know or do?

1. If you have a moment, checking whether your Confluence and Jira user profiles contain a valid email address would be a good thing to do. If they don't, you can fill them in.

2. We are not moving any Confluence personal spaces (including ours) when we migrate. We're also not shutting down the old system immediately, but it will be made read-only once we migrate content. Anything in personal spaces you care about of course, get it saved off sooner, not later.

The plan we have broadly speaking is:

1 Move Confluence MEMBER space and get member support access management working in the Cloud.
2 Migrate the rest of the wiki and make the old site read-only.
3 Install the topic research rewrite for requests to the old wiki.
4 Migrate Jira and member support

We're going to start pretty quickly on 1 and I hope to be past step 2 by end of the month and planning the schedule for 3 and then 4.

When the most disruptive steps occur I will be sending more specific communication or updates. We will do our best to make it as painless as we can, but I don't foresee anything short of "fairly painful" being on the menu. If you have questions they're fine to raise on the users list.

-- Scott