You are reviewing the Engine Performance Logs in Production for a single application that has been live for six months. This application experiences…

The key symptoms described—a high Work Queue - Java Work Queue Size and slow execution of unattended process activities—are direct indicators of an overloaded process execution engine tier. The Java work queue represents tasks waiting for an engine thread to become available. A persistently high value means the rate of incoming work exceeds the system's processing capacity. Adding more engine replicas is the standard Appian scaling strategy to increase parallel processing throughput for background activities, thereby distributing the load and reducing the queue size and execution times.

B. Optimize slow-performing user interfaces.

This addresses front-end performance. The core issue identified is with back-end, unattended processing, not user-facing interface rendering.

C. Add more application servers.

This scales the web tier to handle more concurrent user sessions and HTTP requests, which is not the bottleneck indicated by the engine work queue.

D. Add execution and analytics shards.

Sharding is a data partitioning strategy for managing extremely large data volumes, not for resolving processing throughput bottlenecks under normal sustained load.

---

1. Appian Documentation

System Administration Section

Engine Performance Logs: The Work Queue - Java Work Queue Size metric is described as representing the number of tasks waiting for an engine thread. The documentation states

"A high value for this metric indicates that the engine is receiving more work than it can process in a timely manner." This directly supports the diagnosis of an engine bottleneck. (Found in Appian 23.4 Documentation under System Administration > Health Check > Log Analysis > Engine Performance Logs).

2. Appian Documentation

High Availability and Disaster Recovery Guide

Scaling Appian Section: This guide outlines the architecture and scaling strategies. It specifies that to increase processing capacity for process models

you should add replicas of the execution and analytics engines. This distributes the processing load across multiple servers

directly addressing the scenario's throughput problem. (Found in Appian 23.4 Documentation under System Administration > High Availability and Disaster Recovery > Scaling Appian).