E2E tests (run in CI) are not cleaning up server and project artifacts

[displague]

What happened?

Servers created through E2E testing are not being deleted after the tests complete.

The project name is the same in each test so it is difficult to say which test triggered the project (and to identify if the devices were orphaned as a result of a failed run).

It is possible that an attempt was made to delete the servers and this failed (on the API side). This would block the deletion of the project.

How can we reproduce it?

Ensure that deletion is always carried out on the servers in the project and the project.

Create test servers with a deletion value so that they are automatically reaped.

CLI version (metal --version):

[displague]

Here is how that is done in TF:

• example: https://github.com/equinix/terraform-provider-equinix/blob/main/equinix/resource_metal_device_acc_test.go#L671
• field use: https://github.com/equinix/terraform-provider-equinix/blob/4825cd259b01bff64074cd1bdf78b6ad04f7a8cb/equinix/resource_metal_device.go#L518-L524

[aayushrangwala]

Additionally:

• Each project, device or any resource should have a random generated uuid append ex: metal-cli-project-xxxxx
• Need to add t.Helper() in each test helper in helper.go
• Use t.Cleanup() for declaring the cleanups of the resources created. Suggested here chore: Migrate gateways command from packngo to metal-go client #376 (comment)
• Each resource created, should have a cleanup method called in the test

[aayushrangwala]

/assign @aayushrangwala

[ctreatma]

E2E test runs are still leaving resources behind, even after #405 was merged (and it should be noted that they're also taking far longer to complete).