Understanding DataHub Release Versioning and Upgrade Paths

Original Slack Thread

Hello friends :wave: . We’re trying to understand DataHub release versioning a bit better, and the implications they have when upgrading our deployment:
• DataHub released version strings are “semver like”, but recent patch releases have breaking changes. So presumably DataHub versioning is “semver like” but not quite?
• Is there a documented/supported upgrade path? What’s safe? should we always upgrade serially? (i.e. from 0.10.3 to to 0.10.4, then to 0.10.5) or is it safe to skip some? if so, how many? only patch versions?

apologies if this has been asked before, searching docs and this Slack yield no results for me

<@U05CJD391ND> might be able to speak to this!

<@U04UKA5L5LK> would you mind chiming in on this one? :pray:

Hi <@U0478TGR1M3>! Sorry for not writing back earlier. Good news is that it is safe to skip some! Feel free to upgrade to the latest version you’d like to use :+1:

:wave: Indy, :thankyou1: for chiming in. How many is “some”? would it be safe to upgrade from 0.10.0 to 0.10.5 for example? are we talking about patch releases only? or from minor to minor+2 also (so from 0.8.X to 0.10.X ). A we're not sure answer is also fine, at this point we’re just trying to gauge risk when upgrading, and take adequate precautions if we don’t know the potential outcome of the upgrade.

We have no known issues with upgrading from any version to any version, major or minor! Up until now, the upgrade process has been built cumulatively so you should be able to do an upgrade with no additional risk due to the number of versions being skipped.

thank you!