I'm looking for this concept for Jira too... currently trying to use pgloader, but based on these comments, maybe I should stop.

Our issue is our instances are so big that the XML backup tools don't even work. It's not a matter of things taking days, it's that they simply don't work at scale. We will likely have solved the issue of migration between DBs by the time any Atlassian solution gets developed but the pain and that this involves makes us believe that it should be a built in feature.

Update here after some more work has been done trying to get it to migrate, I'd echo the sentiment expressed by @Jason Smith.  pgloader has proven to be extremely finicky and has a slew of issues that seemingly crop up at random and simply make the conversion fail, making it a solution that would be difficult to recommend for a production deployment.

We have similarly wasted tens of hours with pgloader - there always seems to be some problem at the end of it.

+1 Here.  The fact that this functionality exists for Bitbucket implies that it should be possible to roll out across the suite.  It would take some work though.  Working on doing this with PGLoader with limited success , there are some tables that are clearly not migrating correctly and debugging it is very long and tedious.  Given the near impossibility of generating a full site backup when the site is large, and the fact that there is no other way to migrate global settings, this is something that is definitely needed.

@Hendrik:

the problems you describe are only, if you doing space exports and imports. If you do complete xml backups and restore, these problems don't exist. 

But with xml migration, there are many other problems:

• it takes very long on big instances, which will cause very long downtimes
• on big instances, imports will run into errors, often after hours of importing
• many different problems can occur, which requires to modify the source database or the xml file itself.

did anyone had luck on migrating on database level with external tools like pgloader or ESF Database Migration Toolkit?

This feature would be awesome. Currently there is no good solution for a clean migration. With XML-backup there are the following problems:

• The pages get a new page ID in the new instance. So many external links are broken.
• If Confluence is linked with JIRA, the links between a page and an issue are broken and must be corrected (manually or via database).
• Data from AddOns can not be migrated at all.

"Customers with huge datasets usually end up by developing their own solutions when upon the need of migrating to a different database, which turns into a project at their side that has costs and may turn into frustration using our tools."

This is precisely my current frustration.  Trying to get a Confluence instance with a 32GB BODYCONTENT table to perform a XML backup is simply not possible.  Using tools like pgloader for MySQL -> PostgreSQL still requires DBA knowledge as it is not a clean conversion without a lot of tweaking (which I still have not gotten through).

Bitbucket and Fisheye/Crucible have had this option for a while... would really be nice to see some unity across the apps and have this in Confluence and JIRA.