Create Automtion Cost Optimization Guide

[beyhan]

Collect ideas here for the cost optimization guide:

• Consider using cheaper regions: https://cloudprice.net/gcp/regions?sortField=AveragePrice&sortOrder=true
• Consolidate resources into the same region when possible to avoid data transfer costs between regions. E.g. buckets should be located in the same region as the compute instances.
• Use bosh-lite for test environments where possible
• Automate clean up of orphaned resources like disk, images, VMs... (e.g. in Create Automtion Cost Optimization Guide #1185 (comment))
• Use cheaper and smaller machine types for test environments: https://cloud.google.com/compute/all-pricing?hl=en
• If possible make test environments ephemeral or paused
• Remove/consolidate test cases/environments
• Trigger pipelines/jobs less frequently
• Tag your IaaS resources with BOSH tags for easier discovery with an ops file like:

- type: replace
  path: /tags?
  value:
    <tag-name>: <tag-value>

[peterhaochen47]

Sharing an idea: check if you have a lot of old "compute images" in GCP. You can see those under the "Compute Engine" => "Storage" => "Image" on GCP UI (we had thousands). And you could mass delete them this way:

gcloud compute images list --filter="creationTimestamp<'2022-01-01'" --format="value(name)" | xargs gcloud compute images delete --quiet

compute disks is easier because there is a "In use by" column, and you can tell if the disks are used.