This is a tentative to address test against alpha server making the pipeline fail. The master of server seems to occasionally contain WIP code (ie. time-series collections) and not being always 100% stable.

The idea is to have green pipelines for PR that can be released rather than having failures that can be ignored.

We should still be able to be informed for failures against alpha versions of the server.

We could run tests as a scheduled job (or triggered directly by a new alpha build of server?) from master and get notified with a message on the new-shell channel / team email in case of failures.