Automated scheduling of creation of shared ROSA cluster and another end of business reaper

[kami619]

Adding two new workflow files:

• rosa-cluster-auto-provision-on-schedule
• rosa-cluster-auto-delete-on-schedule

Setting static values for cluster creation, as below:

• CLUSTER_NAME: shared-crossdc
• COMPUTE_MACHINE_TYPE: m5.xlarge
• MULTI_AZ: false
• REPLICAS: 2

and a Rosa cluster reaper scripts, which takes input of a list of "ready" state clusters

• rosa_cluster_reaper.sh

My first attempt at GH workflow's so if something's amiss, please correct it :)

Open question, there could be cluster's which could be in other states other than ready, do we want any alert on it ? Another question is, do we want any email notification after the job is complete, alternative would be to personally tune the GH Workflow notifications to report on this workflow, reducing the email clutter.

[stianst]

@kami619 running out a bit of time before leaving for PTO, and not sure I explained correctly what I proposed, so will try to explain here:

• rosa-cluster-create.yml when executed by default should create the default shared cluster we want
• rosa-cluster-delete.yml should have an option to delete all clusters directly

Now to schedule the above you can either do:

1. Add schedule directly to those workflows
2. Add schedule in separate workflows

Personally I like separating the schedule from the workflow, so leaning towards 2. As you can see in https://github.com/keycloak/keycloak-quickstarts/blob/latest/.github/workflows/schedule.yml it's fairly simple to create a schedule workflow that is only used to execute another workflow. gh workflow run also allows passing inputs to the workflow it's running, so if we don't want to make delete all clusters the default in rosa-cluster-delete.yml it can enable this option when running that workflow.

[kami619]

@stianst and @mhajas I think I have addressed all the review comments. Please review it one more time. Let me know if something is still missing. I will try to run this out of my fork, to share an actual outcome.

[mhajas]

Thanks @kami619 for the changes. I added two more comments.

[kami619]

@mhajas last round 🥊 :) review 🤞🏼

[mhajas]

Thank you for the changes @kami619, LGTM!

I updated the README.md file a little bit because the items were wrongly numbered.