UiPath installation and setup on AWS

Help
, ,
2

I just answer a semi-related question, which you might find helpful

And if you search the forum for “AWS” you’ll find some additional threads on this topic. In short, there is no one solution for all scenario because the topic is so variable. I would highly suggest taking the time to familiarize yourself with AWS and UiPath Infrastructure using the available resources from both Amazon and UiPath, but other content creators (Books, Videos, Blogs, Learning Courses) to prevent any headaches down the road.

  1. How to install Orchestrator and SQL on cloud?

The most basic, provision an EC2 Windows instance with SQL Server, and manually run through the installation. But it will largely depend on your requirements as an example, do you want your SQL Server to be separate from your Orchestrator? If so, then you’ll want to consider one EC2 for Orchestrator and another just for SQL Server, or potentially use RDS instead of EC2 for SQL Server (This will depend on how much you want to manage the host (EC2), OS (Windows), and optimize the SQL Server service vs RDS, you don’t worry about the host or optimizing SQL Server, you just use the service.

  1. Where is BOT developed, on cloud or on local system?
  2. How to deploy BOTs on cloud?

I’m not sure if you’re referring to the development and deployment of the automated process / NuGet package as a “Bot” or referring to the host and/or UiRobot Service that runs on the host.

If you’re referring to the process you are automating, you develop and publish wherever you have Studio installed either directly to Orchestrator, locally on the file system Studio resides, or to any NuGet repository, this can be on a local workstation or as a remote workstation (EC2/VM). Keeping in mind that AWS only provides Windows Server editions for AMI, you need to use Workspaces or import your own Windows consumer edition and manage the license if you want to use that.

In terms of provisioning a machine to AWS for the Robot to run on, you’d want to look at deploying an EC2 instance, there are many ways to go about this depending on how you want to interact with AWS and how much you want to automate the process of preparing the machine image and provisioning the instance itself. Of course, you could also have a host/robot that exists elsewhere so long as you can initiate jobs on it manually or it can communicate with the Orchestrator.

  1. Does application (intranet) which BOT will automate has to be exposed to internet?

Depending on if you have Orchestrator in the mix or not, you need to have the remote NuGet feeds accessible whether that is off the Internet, or if you have a internal private mirror of the NuGet packages or if Orchestrator is providing that for you.

I would read the documentation for Orchestrator and Robot requirements and become familiar with what you’ll need to expose.

There is also the question of the processes are automating and whether they need access to resources on the Internet. When it comes to AWS and other cloud providers, you could have a Private VPN back into your on-prem network and route all public and private destinated traffic route internally, or communicate with it over the public internet. You could also add a Public NAT Gateway to AWS this way your Orchestrator and Robots can initiate traffic to the Internet, but something from the internet cannot.

Essentially you need some way for the Robots / Processes to reach the applications/services it is going to be interacting with. It’ll just depend on your needs and the requirements for acceptable security risk and management.

I’m happy to try and answer some pointed questions or offer opinions given more details or speak more about our current setup. (If you search my replies, you’ll find some threads with some details on that already).

1 Like