Block deploy/promote/merge if degrade performance too much
Description
With Prometheus, we'll be able to monitor the impact of changes in production. But we should catch problems before they are rolled out to 100% of the production fleet and going further, block rolling out changes further if performance has degraded too much. This might be in the form of blocking deploying to production from a canary or blocking incremental rollout past a certain percentage. Or blocking merging something from master
or develop
to a release branch. Or, if using GitHub Flow, the block merging a topic branch into master
.