add retry on db replica create

[DMW2151]

Implements fix for #906 by adding more permissive retry logic in two places:

• Call to client.Databases.CreateReplica in resourceDigitalOceanDatabaseReplicaCreate now allows 412 as a retryable error.
• Call to GetReplica in waitForDatabaseReplica allows 404 while replica creates. Same logic as currently used in resourceDigitalOceanDatabaseClusterCreate.

With the change, one can run a plan that creates a DB and replica as shown below without running into race condition on replica create:

resource "digitalocean_database_replica" "read-replica" {
  cluster_id = digitalocean_database_cluster.postgres-example.id
  name       = "test-postgres-cluster-replica"
  size       = "db-s-1vcpu-1gb"
  region     = "nyc1"
}

resource "digitalocean_database_cluster" "postgres-example" {
  name       = "test-postgres-cluster"
  engine     = "pg"
  version    = "11"
  size       = "db-s-1vcpu-1gb"
  region     = "nyc1"
  node_count = 1
}

Result from apply:

digitalocean_database_cluster.postgres-example: Creating...
...
digitalocean_database_cluster.postgres-example: Creation complete after 3m34s [id=4201c593-eeef-4cac-9465-fade75eaeaee]
digitalocean_database_replica.read-replica: Creating...
...
digitalocean_database_replica.read-replica: Creation complete after 3m53s [id=4201c593-eeef-4cac-9465-fade75eaeaee/replicas/test-postgres-cluster-replica]

[bentranter]

LGTM!