Hi @codemonkee,
It’s been a long time you posted this. Would you mind sharing for what reason you decided not to use NLB? Is this still a reality? I am setting up an environment on AWS ATM and I thought to use this.
Thanks 
I don’t recall the specifics at the moment, but it had to do with the traffic not being passed correctly when accessed via Route53 to a NLB and to the Orchestrator Node - We didn’t follow through on using an AWS NLB for a couple unrelated reasons.
- We hadn’t purchased the UiPath HAA licensing yet (Which we preferred over rolling our own unsupported Reddis.
- By that time, we had start migrated our UiPath Infrastructure off of AWS for cost savings and into our Nutanix Private Cloud environment.
Other than being hosted on-prem within our Nutanix eco system, our design mostly remains the same. Instead of AWS NLB we are using a pair of F5 Load-balancers and will pool two Orchestrator nodes. (We’re actually going to be setting that up soon as we just got the HAA license a few weeks ago, and started upgrading to 2020.10 earlier this week. We still need to upgrade our Production environment at some smoke testing and then will look at finishing the HAA / Reddis Cluster for each environment.
I haven’t review the files, I imagine they’ll be slight different from how I set things up, but have a look through UiPath’s Infrastructure Github repository, they have a number of different templates for different cloud providers including AWS that might give you some inspiration.
I also believe internally they are working on more automated deployments for customers beyond Azure.
1 Like