terraform-provider-vultr: rate_limit does not seem to be limiting the request rate to Vultr properly preventing the use of -parallelism=24

Vultr has a horribly slow request rate limit at 2/sec. So when using this provider driver combined with other providers or have a node count higher than 2-3 it generates rate limit errors even on terraform plan. You have to end up using terraform plan -parallelism=3 so it even works slowing down your overall deployment considerably. The only workaround I have found is to add a provisioner for a local 10 minute sleep after each vultr server deploy and separately creating a plan file first. However, this is unreliable. provisioner "local-exec" { command = "sleep 600" }