I am playing with an idea of using deployment groups along with tags to delay a release to a server until it is tagged as offline.
Scenario:
- I have a single QA server in my QA deployment group by default it doesn't have the tag 'offline'.
- I have a release pipeline that targets the QA deployment group but only servers that are tagged as 'offline'.
- I have the Auto-redeploy trigger on the QA deployment group release stage to redeploy when new servers are added with correct tags.
So when i trigger a release it will initially not deploy to the QA server but then at a later point in time the QA server will be tagged as 'offline' and i would like the release to then deploy to the server.
I have tested this out and have found that if the initial deploy fails to deploy to any servers the release is treated as a failure and so an Auto-redeploy never happens even when the server has the tag added.
If the deployment group has more than 1 server in it and the initial release deploys to at least one server (so say one already has the 'offline' tag), then any subsequent Auto-redeploys work as expected when a server is tagged as 'offline'.
So my question is how can i make the first scenario work?