Creating a new auto-deploy branch seems to result in a deployment, even if the branch' tests are failing
We ran into this today: we created 13-2-auto-deploy-20200714
, and despite its tests failing we still ended up deploying it to staging and canary. This suggests that whenever we create a new auto-deploy branch, we ignore the state of its tests. This can result in us deploying broken code without even knowing it. Considering the potential impact I think this warrants a DeliveryP1