A parallel run is the part of migration where you're paying for two environments simultaneously -- old system still live, new one proving itself. For an org like NZRT, that window needs to be short. We don't have the budget headroom to sustain double infrastructure indefinitely, and our suppliers aren't going to give us a discount for indecision.
What keeps me focused is the other end of this. The Iteasel stand in a community space, the Nextcloud instance a charitable organisation couldn't have self-hosted without us. If we drag out the parallel run window, we're burning capacity that should be going toward delivery. Every week we're running two stacks is a week we're not standing up the next one for a community that's waiting.
The Migration Planning and Transition Architecture docs have the cutover sequencing. The goal is a clean parallel run window -- long enough to validate, short enough to keep costs contained.
Migration Planning: https://cloud.nzrtnetwork.com/nextcloud/index.php/s/zqgqF8PPAJnpkem
Transition Architectures: https://cloud.nzrtnetwork.com/nextcloud/index.php/s/gcJdZkWe8XCoXki
