Improved pooling of resources

[HakanEngman]

Hi,

Current scheduling options are quite limited (static) when it comes to allocation of resources. We would like a schema to be able to allocate one resource in a pool and execute all tasks in a sequence. 

Adress: Overlapping schemas on the same resource. With a larger pool we could cut the overhead needed to secure robot processing. Also, a crashed/stuck robot would not affect the following schemas. (Security reasons imply usage of different robot accounts for different processes.)

[MariusBalukevic]

• Not able to use active queue and resource pools functionality, because it is not applicable for multiple users. We have to dedicate particular VDI to schedule robots and it cause us lots of scheduling incidents

[OskarVikholm]

Hello,

I requested this a year back as an enhancement but no action was taken. Basically, there are several Blue Prism customers that would like to see added support for using resource pools together with Login Agent. This would save a lot of time and trouble in scheduling of automations, including better utilization of licenses.

Currently, it is only possible to use resource pools in a smart way by not using login agent, i.e. to allow the Windows Account to have access to all systems with Single Sign-On (SSO). The other option to get around this is to build custom scheduling which Blue Prism would not support as part of the license agreement.

In case someone solved this in another good way, please write a comment here.

Thanks

[TomCirone]

​Somewhat related...it would be good to be able to call a specific resource even if it's in a pool.  If there's an application that has an expensive license, we'd only want it on the minimum number of machines.  Then we have to keep that resource outside a pool

[Mike__Lowe]

Hi Hakan, thank you for raising your idea.

Are you able to describe in a little more detail the idea you have raised?

• What exactly do you see changing in the product?
• What is the exact issue that this would overcome?
• How would this change benefit your organisation?

Many thanks, 

Michael Lowe

[HakanEngman]

Right, I'll try to outline a bit more. 

Today we schedule robots with multiple steps and one key thing we need to comply with is the usage of different robot accounts for different runs. 

We do login with different accounts supported by login agent setup, but we need to stipulate the resource we want to run on, not a "available one". 

Would like to have a option where you make a pool out of a number of resources. Priority robots  (could) for example have access to all available resources, less important robots could have a limited number of resources to choose from. 

It would be great if a schema setup would use one resource for the execution, not set per step as of now. ( 1 Schema = 1 Resource ) 

Then next evolvement would be 1 Schema = 1 Resource Pool ( 1 Schema run always use one resource from start to finish in one action )

Basically, the schema would login one robot user and do the work and logout.

As the resources in our case are identical. (Virtual desktops) And I assume that many customers running in public clouds have the same feature with EC2's and similar technologies. 

So, to try to clarify.

Add resources to a pool, make it possible to phase out resources for maintenance( do not accept new schema runs )  One schema should be loosely coupled to a pool, not a static resource. Make all steps execute on one resource, not like now when the first step execute on R1 and step to on R3 and in a random way. )

So, we really want to move away from the fact that all resources in the pool needs to be logged in by one single robot account. 

( Already today you can pull a pool member from commandline, but not from the graphical interface ) 

• We would like to be able to utilize the resources better by running the schemas closer in time. 
  
  
•  We would like to improve resiliency against different problems like "blocked" resources etc. 
•  Make it easier to do maintenance tasks on the resouces. 
•  A higher utilization on the hardware would make the Blue Prism solution more cost effiicient (w.o. impacting revenue for BP... ) 
•  A operations team only need to make sure that there are headroom enough to accommodate the max parallell sessions available in BP.

Please let me know if this needs to be more outlined 

[MelanieGiuliani]

Hi Hakan,

Thank you for providing some additional information on your request. As a result of the new information you've provided, we are moving this idea into the Under Consideration status while we route it through our internal review process. 

We will update you as your idea moves along the lifecycle.

Thank you!
Melanie 

[Stefan__Pappalardo]

Hello Everybody, additionaly I can share one of my support requests created at 2018-12-10 09:58:33. It's quite similar to the idea from Hakan and perhaps will add some additional insights.

#47692 Change logged in user before running business process with scheduler, login agent and resource pool.

Our business processes need a dedicated user for every single process. This is a hard restriction (security policy) we have to deal with. Until now we schedule sessions (using the internal scheduler of Blue Prism Automate) for one virtual machine (vdi), with one logged in user, to run one dedicated process on it. The login is performed by login process (from Blue Prism login agent) once a day to control login agent installed on all runtime resources. "Automate.exe /resource-pc" is started on login (autostart). This scenario leads to many virtual machines idle waiting for the next run. Every machine costs money and is not utilized as efficient as it could / should be.

We tried to use the resource pool feature in conjunction with the scheduler and the login agent. All features are part of the Blue Prism Automate product. But when it comes to a user change (logout & login) the connection between BPA server and runtime resource gets lost which breaks the use of the scheduler.
Another problem is to hit the same resource for user change and run business process in sequence using resource pools.

• Is there a way to configure the login agent to hold the connection between server and runtime resource?
• Are there any plans to add user change prior to run business process in the scheduler?
• Or do you have another idea to solve our needs?

The answer from support was: The best way to raise an enhancement for the product would be via the Portal. We have a form you can complete which ensures you provide all the information the Product Team need to be able to consider the product change you would like and the form goes directly to them. The form can be found here: https://portal.blueprism.com/products/enhancement-request

Meanwhile the idea system was added and there was no migration of the already reported enhancement requests. But I guess I never raised one because I didn't trust this form any more after raising some requests and never ever heard anything about the state.

@MelanieGiulian1: But I can provide a more in depths mail discussion with John Carter about Pools with Login Agent from February 2019 if you need.​​

[robert.nicklin]

Hello Hakan,

Thank you for posting your idea.

Having reviewed the current Blue Prism portfolio roadmap and product strategy I have determined that this enhancement is unlikely to be implemented in the next 12-24 months. As a result, I'm going to update this idea to "Not planned".

Should this position change we will review your idea and update the status accordingly.

Regards,
Rob

[chris.strong]

Hello @HakanEngman 

I believe we have solved your original request with Automation Orchestrator (powered by Next Gen) and the recent additions of Dynamic Worker allocation with process chaining.

I appreciate this idea request was raised a long time ago. It is one of the multiple customer points of feedback which led us to deliver Automation Orchestrator (powered by Next Gen). Our latest enhancements include Automation Flows, Dynamic Worker allocation, Process Chaining, Digital Worker Groups, and Triggers, which provide flexibility, efficiency, and control over automation processes. These capabilities are available to our Blue Prism Enterprise customers with Automation Orchestrator (powered by Next Gen).

Our Automation Flows (which supersede schedules) allows you to create sequences of automations that can be executed in a specific order, and separately define multiple triggers, including Work Queue triggers and time-based triggers. We have clash detection built-in, so that if a digital worker is busy, the next session (process run) would be queued, ready for when a worker becomes available, giving you greater resilience and utilization.

Our Digital Worker Capability Groups (which supersede Resource Pools) enables customers to define and manage groups of digital workers with specific capabilities, ensuring that automations are executed by the most suitable workers.

• A digital worker group is a collection of digital workers with a shared set of capabilities, defined within an environment.
• Within an automation flow, when defining a step (task) that contains an RPA process run, you can choose a digital worker group to run the process instead of selecting an individual digital worker, and the next available online digital worker in the selected group then runs the process.

In addition, our Process Chaining capability enables customers to create sequences of automations that can be executed in a specific order, including login and logout steps.

• A Process Chain allows you to chain together multiple processes in a process run to leverage Login Agent to securely log in to an execution machine while also benefiting from dynamic digital worker allocation. You can create a chain of processes in a process run to dynamically select any appropriate digital worker from a group and ensure that it securely logs in to an execution machine using specific credentials, before then running the required RPA automation. Once the automation is complete, the machine can be cleaned up and logged out, ready for the next automation.
• Tip: Even if the first or second process in the chain fails, the third process will always run to ensure that the machine is correctly cleaned up, logged out, and returned to a state where it is available to be used again.
• For processes in a process chain, the digital worker is reserved by the Automation Orchestrator until the last process in the chain is complete, regardless of whether it was explicitly selected or was allocated from a digital worker group. This ensures that the same digital worker runs each process in the chain, and cannot be allocated to a different session when it completes one process and is waiting to start running the next one in the chain.

This allows for more efficient and streamlined automation workflows because a chain can be assigned to any worker within a customized capability group.

To learn more about Automation Orchestrator (powered by Next Gen) and how it can help your organization streamline and optimize your automation workflows, please visit our online documentation portal at https://docs.blueprism.com/en-US/bundle/next-generation/page/automation-orchestrator.htm.

Related ideas where I've shared the same update:

• Improved pooling of resources
• Resource Pools & Login Agent & Multiple Processes
• Login Agent with Resource Pools
• Scheduling Login Agent and Main Process in Resource Pool
• Improved pooling of resources

I am really excited to see the impact that these capabilities will have on our customers' automation initiatives and look forward to continuing to innovate and deliver value to our community.

Kind regards

Chris Strong

SS&C Blue Prism – Senior Product Manager (Next Gen RPA)

[Otto_Mation]

This idea has been updated by the SS&C Blue Prism Product Management Team