Hardware requirements

Action Center can be installed using multiple node deployments. Depending on the type of node deployment you are using, you need to consult different hardware requirements:

• Single node deployments - single node Action Center connected to a single node Orchestrator instance
• Multi-node deployments - we recommend the following multi-node deployments:
  • In a multi-node deployment scenario, the Orchestrator is set up behind a load balancer, while a single instance of Action Center is installed on a separate server. Action Center is configured to point to the URL of the Orchestrator's Load Balancer.
  • Another type of multi-node deployment involves setting up the Orchestrator behind a load balancer with multiple nodes. In this configuration, Action Center is installed on each Orchestrator node, and each instance of Action Center points to the Orchestrator's Load Balancer URL.

Action Center can scale up to a high number of concurrent users if the connected Orchestrator can handle load as per the requirements documented here.

Single node deployments

Single-node hardware requirements

Based on an in-house load test conducted within a sandboxed environment, we provide the following hardware requirements recommendations. These recommendations are based on the sample data mentioned in the section below.

Note:

Customers need to derive hardware size based on their business requirements (average payload size, load expected, concurrent users, etc).

Max. Concurrent Users	CPU Cores (min 2GHz)	RAM (GB)
4,000	2	4
12,000	4	4

Test Setup and Sample Data

To test the performance for a single node Action Center connected to a single node Orchestrator instance, the below setup and sample data was used:

• 3 hours of test time
• Form Data payload having a size of 5000 bytes
• 28,000 actions created
• 17 different API endpoints executions
• Simulated concurrent users

Multi-node deployments

Multi-node hardware requirements

Based on our performance tests, we recommend the maximum numbers below:

• 150,000 attended robots connected to Orchestrator and running jobs
• 10,000 Action Center users processing actions
• 3,000 unattended robots running jobs (creating actions and resuming after the actions completion)

Environment Configuration

To ensure seamless performance for the above scenario, we recommend setting the environment below.

Instance	No. of Deployment Nodes	Azure VM	vCPU Cores	Frequency (GHz)	RAM (GB)
Action Center	3	B2s	2	2.0	4
Orchestrator	10	F16	16	2.0	32
SQL Server	1	F32	32	2.0	64

Drive Storage

Allocate a drive for each of the following content:

Stored Content	Drive Capacity
Database	1TB
Temporary database	1TB
Transactional logs	1TB

Database Setup

Product	Configuration
Redis Enterprise HA	CentOS 8 CPU Cores 2.0 GHz minimum frequency 16GB RAM
Bucket storage VM	Standard L32s_v2 Ultra Disc 4TB 900MB/s throughput

Test Setup and Sample Data

We generated the hardware requirements, along with additional setups and configurations, by conducting performance tests that simulate a large load on Action Center and Orchestrator. The tests were performed using the following sample data:

• 10,000 concurrent Action Center users
• 240,000 actions
  • 60% document validation actions
  • 40% form actions
• Payload per one form action:
  • Form layout—5KB
  • Form data—5KB
  • 10 storage files x 100KB
• Payload per one document validation action:
  • One PDF file x 150KB