Q&A: Deployment templates

• Q: For the multi-node HA-ready production profile, the processor requirement for the Complete product selection is 96 (v-)CPU/cores. Is this requirement inclusive of server and agent nodes, or does it refer to server or agent nodes?

• A: Our documentation indicates that the multi-node HA-ready production profile requires at least 3 server nodes with minimum CPU/RAM per node. As long as this condition is met, the remaining CPU/RAM requirements can be spread over either additional server or agent nodes that also meet the minimum CPU/RAM requirements.

• Q: Is a GPU required for Task Mining?

• A: No, GPU is not required for Task Mining.

• Q: What does node fault tolerance in the UiPath® Automation Suite Install Sizing Calculator refer to? Does it cover agent nodes, only server nodes, or both?

• R: La calculadora de tamaño de instalación de UiPath® Automation Suite basa la recomendación en la tolerancia a fallos para que siempre haya un mínimo de 3 nodos de servidor disponibles en el peor de los casos. Puedes aumentar el número de servidores que pueden dejar de funcionar sin que la implementación falle.

• Q: Can I use a private domain in Amazon Route 53?

• A: Yes, private hosted zones are supported.

• Q: Do I need to specify the DNS records in the cloud solution (e.g., AWS) as well? Can I use an on-premises DNS solution with an Automation Suite AWS deployment? For example, in the case of a URL the end-user accesses, does the record need to be resolvable in the cloud or on-premises DNS?

• A: You can use non-Amazon Route 53 DNS infrastructure. However, the DNS must be resolvable in the VPC where you perform the deployment.

• Q: Which CNI plugin is used as part of Kubernetes? For example, when vendor-supplied, the Kubernetes solution supports integration with VPC/VNET. If using AWS Kubernetes, whenever the vendor creates and deploys workload, pods get IP addresses from the VPC itself. That means the workload can be accessed by AWS or Azure services.

• A: We use Cilium CNI. The Kubernetes distribution comes with 3 supported CNIs, but not the AWS-specific one.

• Q: Do you customize Service IP and POD IP ranges?

• A: We currently do not customize Service IP and POD IP ranges. The default CIDR is 10.42.0.0/16.

• Q: When deploying Automation Suite to AWS into an existing VPC, is it possible to use TerraForm instead of CloudFormation?

• A: No, it is not possible. We use the native Infrastructure as Code tool from the cloud provider (CloudFormation for AWS, ARM for Azure).

• P:En cuanto a la pregunta anterior, si mi estándar para IaaC es Terraform, ¿significa que tengo que desarrollar el código de Terraform?

• A: If you would like to deploy Automation Suite manually (not using the existing AWS or Azure templates), then yes, you would need to develop the Terraform code. Alternatively, you can use the existing CloudFormation template for deployment or as a base to assist in creating the Terraform code.

• Q: How can I estimate the number of specialized nodes required by Task Mining?

• A: Task Mining requires/supports 1 dedicated node out of the box.

• Q: Can I use Graviton-based processes for the deployment?

• A: No, the Kubernetes distribution does not support ARM.

• Q: Can Automation Suite be hosted on AMD or only on Intel processors?

• A: Either is fine as long as you meet the architecture requirements.

• Q: Can I deploy Automation Suite templates on Ubuntu and CentOS?

• A: No, Ubuntu and CentOS are not supported.

• P: Para una implementación de AWS, si un EFS se implementa automáticamente, ¿cuánto espacio reserva?

• A: EFS does not come with pre-provisioned size. It is pay-as-you-use. EFS will hold all backups for all disks across the cluster, which means it will use 6 TB of data, the sum of all the nodes storage data disk space.

• Q: Does Azure deployment deploy an NFS solution, as AWS does? If it does, how much space does it reserve?

• R: NFS es un protocolo que tiene múltiples implementaciones posibles. AWS Elastic File System es una de esas implementaciones. El servidor NFS es otra implementación. Azure tiene una implementación diferente llamada Azure Files. Si quieres tener una copia de seguridad, necesitas uno de estos elementos de infraestructura.

• P: ¿Cuál es el proveedor de servicios en la nube preferido: Azure, AWS o GCP?

• R: No tenemos un proveedor de servicios en la nube preferido. Los admitimos todos.

• Q: If Autoscaling is enabled, does the node require customized images?

• R: El autoescalado utiliza tu imagen de RedHat, sobre la cual se instalará el software UiPath®. No tenemos datos claros ahora, pero el inicio en frío está en el rango de 5-10 minutos.

• Q: Can I use the Standard edition of Microsoft SQL Server?

• A: We support Microsoft SQL Server 2016, 2017, and 2019 Standard and Enterprise editions. For more on SQL requirements, see our documentation.

• Q: Is the the total vCPU requirement of 96 cores required from the beginning or can it be part of Autoscaling, which will add or remove nodes depending on the resource need?

• A: 96 cores are required to power up the system. Subsequently, you can employ different scaling strategies. However, you must not go under that number.

• Q: Do I select the RHEL image or is it prescribed in the AWS or Azure templates? If I must select it, what RHEL version should I use when deploying via templates?

• R: La implementación utiliza una imagen RHEL básica (que se determina en tiempo de ejecución) y crea una plantilla de grupo de autoescalado que se aplica cuando se aprovisiona la máquina virtual. Esto significa que todo el software se instala cuando se crea una nueva instancia EC2. Las plantillas utilizan RHEL 8.2 a partir de ahora.

• Q: Is there any need for VMs other than for the Kubernetes cluster?

• R: Para ejecutar Automation Suite, no se requieren máquinas virtuales adicionales fuera del clúster. Sin embargo, si no tienes máquinas existentes para ejecutar UiPath® Robot, es probable que estas sean necesarias, además de cualquier máquina virtual o escritorio de desarrollador para UiPath Studio (IDE).

• Q: Will there be any integration with CloudWatch or Azure Monitor for Kubernetes logs?

• A: As of now, there is no integration to push logs to CloudWatch.

• Q: Can Automation Suite use open-source database solutions, such as MySQL or PostgreSQL?

• A: The Microsoft SQL Server requirement is independent of the cloud where the Automation Suite is deployed.