Robotic Process Automation

How do I scale my RPA bot architecture from 5 to 50 unattended robots efficiently?

MA Asked by Mark Robertson · 14-04-2024
0 upvotes 12,493 views 0 comments
The question

We have successfully automated five back-office processes using UiPath, but as we look to scale to 50 bots, we are hitting performance bottlenecks. Our current infrastructure is struggling with concurrent executions and database locks. What are the best practices for setting up a high-availability Orchestrator environment and managing machine templates to ensure smooth scaling? 

3 answers

0
HE
Answered on 16-04-2024

To scale effectively, you must move toward a virtualized, elastic infrastructure. Ensure your Orchestrator is in a Multi-node setup with a Load Balancer to distribute the traffic. Instead of static machine assignments, use Machine Templates and Dynamic Slots; this allows the Orchestrator to allocate any available VM to a pending job. Also, look into your SQL Server performance; as you scale bots, the logging frequency increases. Consider adjusting your log levels to "Information" or "Warning" only for production to reduce DB I/O. Finally, implementing a robust "Hub and Spoke" model for your Center of Excellence (CoE) will help manage the governance of those 50 bots without causing a management nightmare. 

0
JA
Answered on 18-04-2024

Are you utilizing a persistent or non-persistent VDI setup for your robot machines during this scaling phase?

MA 19-04-2024

Jason, we are currently on persistent VMs, but we’re seeing significant costs. Would non-persistent be better? Definitely. Non-persistent VDIs combined with a "floating robot" license model in UiPath can save you about 30% on infrastructure costs. It ensures that when a bot finishes a task, the VM is wiped and ready for the next one, preventing profile bloat and session hanging which often slows down large-scale RPA deployments.

0
M
Answered on 20-04-2024

Don't forget about monitoring. With 50 bots, you need a dashboard like Kibana or Tableau to track heartbeats and success rates in real-time. 

MA 21-04-2024

I agree with Megan. We used a custom ELK stack for our bot monitoring and it made identifying which of our 40 bots were failing much faster.

Share your thoughts

Your email address will not be published. Required fields are marked (*)

Professional Counselling Session

Still have questions?
Schedule a free counselling session

Our experts are ready to help you with any questions about courses, admissions, or career paths. Get personalized guidance from industry professionals.

Request a Call Back

Search Online

We Accept

We Accept

Follow Us

"PMI®", "PMBOK®", "PMP®", "CAPM®" and "PMI-ACP®" are registered marks of the Project Management Institute, Inc. | "CSM", "CST" are Registered Trade Marks of The Scrum Alliance, USA. | COBIT® is a trademark of ISACA® registered in the United States and other countries.

Book Free Session