Developers hate it, users need it: Backward compatibility. Adding new features is great, but not gaining adoption of those, because your users are still stuck on the 2011 release that is riddled with bugs, will neither help your product, your users or your developers or support folks. Making sure that upgrades are smooth and backward compatible will help your users to stay up to date.
This talk outlines the general Elasticsearch strategy when it comes to backward compability. After deprecating a feature in one major version, the user needs to be made ware of this before it can be removed or replaced. We will take a look at the code required for these steps and how the user gets some help, so he does not stumble unprepared into a deprecation.
We will also dive a little bit deeper into the migration from one date parsing and processing library core java time and talk about the caveats and the process that had been done to support this.