Context:
From the back-end senate (https://confluence.camunda.com/display/CO/2020-03-25+The+back-end+Senate+-+March+II):

• Writing an upgrade script is a real pain. Instead of using the bulk API stuff and adding the reports/collections/alerts etc. manually I would like to reuse the API that we already have.
• Does it make sense to have an Optimize running for the upgrade scripts that check that the reports work with the running Optimize and not just checking everything is fine in ElasticSearch structure?
• it would be nice

AT:

• need to be defined

Hint:
let's think about having a dedicated test extension to start the old version of Optimize in a docker container and then perform some requests against the old version. After, the migration is performed and in the end we use the new Optimize version api to check if the result is correct.

This would have the following benefits:

• it's easier and more understandable to write tests
• we don't need the complicated optimize data upgrade pipeline any longer
• we are actually testing e2e that the migration works within our upgrade module.