Refactor queue to handle more queue entries

@yas, I've added support for creating multiple queues using a deriver. I've only added the deriver to the AWS, K8s, Openstack, Terraform, and VMware. I don't think this is needed for cloud_cluster or cloud_cluster_worker.

@yas, setting back to needs work. I need to make queue changes to K8sCloudStorePlugin.php as well. Those were caught during Unit Testing.

@yas, please review the refactoring. I've extracted the logic that adds items to the queue into two helper methods.

getResourceQueueName and updateResourceQueue.

I've also updated the following

• .tugboat/config.yml
• cfn/nested/ec2/ami/template.yaml

@yas, all the requested changes have been made.

@yas, please review the updated.

@yas I've added additional functionality to this patch for queue refactoring.

1. Added a new drush command drush cloud:queue:cleanup. Example usage of the command
drush cloud:queue:cleanup --hours=5 --queue_names=queue1,queue2 . Delete items from a comma delimited list of queues that are older than 5 hours. This lets you delete from any queue of your choosing, even queues created by other modules.
drush cloud:queue:cleanup. This is the default usage and will delete all queue items older than 1 hour. It only deletes from queues created by cloud service providers. Queues created by other modules are ignored.

2. Added try/catch around all queue processing. This way, if there is an exception, that particular queue item will be booted out of the queue. By doing so, we do not have lingering queue items that can potentially error out.

3. When deleting a cloud service provider, delete all their queue items. This will help clean up the table a bit. I've done this in hook_cloud_config_predelete(). Doing so won't cause any CloudConfigPluginException errors to be thrown mistaken.

4. In the processItem of the Queue Worker plugins, I've abstracted the common function that checks if a service method exists. I'm also checking to see if setting the CloudContext will cause any CloudConfigPluginException errors.

Finally, I did not hide the configuration for *_queue_limit variable yet. We haven't really testing if count=1 is truly desired yet. If we want to do this, we can create another issue.