app-autoscaler base manifest switch

[risicle]

What

https://www.pivotaltracker.com/story/show/184549339

This involves moving almost all jobs out into their own instance groups and changing a lot of naming to be based on ((deployment_name)), meaning quite a bit of hostname adaptation etc.

It also currently points manifests/app-autoscaler/upstream at our own fork until https://github.com/cloudfoundry/app-autoscaler-release/pull/1653 or a similar fix makes it upstream. Left as a draft until that happens (or we abandon all hope).

(This also means this is likely to bump to the latest autoscaler version once that happens so we can include that change)

How to review

Deploy to a dev env, see the autoscaler acceptance tests passing.

Check the dev env's alertmanager (https://alertmanager-1.<env-domain>) to ensure AppAutoscalingIsNotScalingSingleProc and AppAutoscalingIsNotScalingMultiProc alerts aren't going off.

---

🚨⚠️ Please do not merge this pull request via the GitHub UI ⚠️🚨

[kr8n3r]

is this still relevant?