chore(ci): move from container to VM on cncf hosted runners

[koksay]

We noticed that the workflow might have performance issues running on a container on the CNCF-hosted runner. This PR moves it to a Virtual Machine.

Below is the list of available VM based hosted runners: (you may need to enable them in your organization's settings)

# amd64
oracle-vm-2cpu-8gb-x86-64
oracle-vm-4cpu-16gb-x86-64
oracle-vm-8cpu-32gb-x86-64
oracle-vm-16cpu-64gb-x86-64
oracle-vm-24cpu-96gb-x86-64
oracle-vm-32cpu-128gb-x86-64

# arm64
oracle-vm-2cpu-8gb-arm64
oracle-vm-4cpu-16gb-arm64
oracle-vm-8cpu-32gb-arm64
oracle-vm-16cpu-64gb-arm64
oracle-vm-24cpu-96gb-arm64
oracle-vm-32cpu-128gb-arm64

[jaydeluca]

We noticed that the workflow might have performance issues running on a container on the CNCF-hosted runner.

out of curiosity, where did you see this?

[koksay]

We get CPU throttling alerts on the Kubernetes cluster running those jobs.

[otelbot]

Thank you for your contribution @koksay! 🎉 We would like to hear from you about your experience contributing to OpenTelemetry by taking a few minutes to fill out this survey.