Document Automation - scaling requirements

This document outlines the approach to scaling Document Automation, including determining infrastructure requirements and configuring essential platform components.

Overview

Document Automation utilizes Bot Runners to perform extraction tasks. These Bot Runners are key to executing workflows, with each handling a portion of the workload. The platform’s orchestration mechanism intelligently distributes tasks across available Bot Runners, ensuring that none remain idle while others are overloaded. As each Bot Runner completes a task, it is automatically assigned a new one, maintaining balanced utilization.

By increasing the number of Bot Runners, you can accommodate varying document volumes, providing high throughput and flexibility to meet business needs—a process known as horizontal scaling. In many scenarios, system capacity scales nearly linearly with additional Bot Runners, but factors like network bandwidth, database access speeds, and I/O operations might introduce diminishing returns at higher scale levels

Determine required resources

To choose the best scaling approach, begin by calculating your required daily processing volume. A day refers to a full 24-hour cycle. If your bot operational hours are shorter (for example, an 8-hour workday), adjust your Bot Runner throughput accordingly. For example, while one Bot Runner for a particular document type might process 2,400 pages per 24 hours, it could be expected to only handle up to 800 pages within 8 hours. Plan resources based on your actual use cases and processing window.