convert from provisioned_capacity to provisioned_vcpus #222
Conversation
Use new plan types. Remove support for spend limits. Add support for provisioned clusters. Adjust to use *int64 for usage limit fields rather than int64.
Previously, the TF provider raised an error when the primary attribute was
specified in multiple regions, even when false in all but one region, e.g.:
regions = [
{
name = "eu-west-1"
primary = true
},
{
name = "eu-west2"
primary = false
},
{
name = "ap-south-1"
primary = false
},
]
This PR fixes that by adding some custom validation logic rather than using
the built-in config validators.
Fixes: CC-27160
Make the "plan" attribute on the "cluster" resource writable so that it can be set by users who intend to upgrade/downgrade the plan. Keep the attribute as computed so that the system will set the value if it is not provided by the user. Release note: Support upgrade of plans via changes to the "plan" attribute on the "cluster" resource.
Rename the `provisioned_capacity` field to `provisioned_vcpus` and update all tests and examples to use the vCPUs unit rather than request units by dividing by 500. This is part of the overall migration to vCPUs for Cloud 2.0.
|
Similar to the SDK PR, I will hard reset the IMPORTANT: All but the last commit are older commits that are being reapplied to a new base. All of the older commits have already been reviewed. While I did need to make a few tweaks to those to get past some merge conflicts, there are no material changes. I'd advise only reviewing the last commit. |
|
Also, I won't merge this PR until all of the required API changes are in production (it seems we're still pretty far behind in our rollouts). |
|
I will hard reset the current version of the Cloud 2.0 TF provider to this commit. Only Private Preview customers have been using this by manually cloning and building this branch. They will need to switch any config they have to use |
Rename the
provisioned_capacityfield toprovisioned_vcpusand update alltests and examples to use the vCPUs unit rather than request units by dividing
by 500. This is part of the overall migration to vCPUs for Cloud 2.0.
Commit checklist
make generate)