Enterprise Troubleshooting Guide for TestComplete in CI/CD Pipelines

Details: Category: Testing Frameworks; By Mindful Chase; 27.Jul; Hits: 7

TestComplete is a widely used automated UI testing tool for desktop, web, and mobile applications. While its ease of use and support for multiple scripting languages make it attractive, enterprises often face obscure, environment-specific, and high-impact issues when scaling TestComplete in CI/CD pipelines. This article dives into some of the least documented yet highly disruptive issues such as object recognition failures, inconsistent test results across agents, and memory-related crashes. We'll explore root causes, architectural dependencies, and systemic fixes that can drastically improve the reliability and maintainability of TestComplete-based testing infrastructures in large-scale environments.

Mindful Chase

Writing Code, Writing Stories

tbd

Experience

tbd

More to Explore

Understanding the Scope and Challenges of TestComplete

How TestComplete Fits into Enterprise QA Pipelines

TestComplete is typically used for end-to-end functional testing across desktop and web apps. It integrates with Jenkins, Azure DevOps, and other CI tools. Enterprises often run hundreds of test suites across virtual machines, making stability and repeatability critical.

Core Issues in Enterprise Environments

Object recognition failures on virtual/remote machines
Inconsistent results between test agents
TestComplete hanging or crashing during long runs
Memory leaks and resource exhaustion
License-related execution failures in concurrent pipelines

Diagnosing Object Recognition Failures

Symptom

TestComplete cannot find or interact with UI objects during test execution, even though the object exists when inspected manually.

Root Causes

Incorrect or outdated NameMapping configurations
DPI scaling or display resolution differences
Virtual environment rendering delays
Window layering or minimized windows

Resolution Strategy

#1 - Validate NameMapping in the environment where failure occurs
#2 - Disable "Hide the desktop while running tests" in project settings
#3 - Set screen resolution and scaling via group policy or scripts
#4 - Use property-based identification instead of image-based

Test Inconsistencies Across Agents

Symptom

Tests pass locally but fail intermittently on build agents.

Key Areas to Investigate

Different OS patch levels or user profiles
Screen lock or session timeout during execution
TestComplete version mismatches

Mitigation Strategy

# Ensure identical agent images (use golden images or Docker if possible)
# Keep session active with PowerShell:
  Add-Type -AssemblyName System.Windows.Forms;
  [System.Windows.Forms.SendKeys]::SendWait("{SCROLLLOCK 2}")
# Synchronize execution timeouts in CI
# Pin all agents to the same TestComplete patch version

Handling Memory Leaks and Crashes

Symptoms

TestComplete freezes or crashes mid-execution
Windows reports application memory usage spikes

Diagnosis Approach

# Use Windows Performance Monitor to track memory and handle count
# Enable TestComplete logs and check for "Exception Code: 0xC0000005"
# Monitor for large image or video logs bloating output folders

Best Practices to Avoid Crashes

Split long-running tests into smaller suites
Disable unnecessary logging (screen captures, video logs)
Restart TestComplete and the VM between test batches

Licensing and Execution Issues

Typical Problem

Concurrent test runs fail due to license server overload or denials.

Architectural Recommendations

Use Floating User Licenses instead of Node-Locked
Implement license usage monitoring with SmartBear License Manager
Stagger test job execution via queue management

Automation and CI/CD Best Practices

CI Agent Hardening

# Disable Windows Defender real-time scanning on TestComplete folders
# Pre-load dependencies before execution starts
# Run TestComplete with elevated privileges and set process priority

Execution Scripting Example

TestComplete.exe /run /project:"MySuite" /exportlog:"logs/result.mht" /exit
if %errorlevel% NEQ 0 exit /b %errorlevel%

Conclusion

Enterprises using TestComplete at scale must look beyond surface-level scripting and address systemic issues across CI agents, environments, and infrastructure. By proactively managing environment configurations, isolating memory leaks, and structuring test suites modularly, teams can dramatically improve the reliability of automated testing pipelines. Proper planning around licensing, object recognition, and platform-specific tuning is essential for consistent, reproducible results in TestComplete environments.

FAQs

1. Why does TestComplete fail to find objects on remote machines?

Display settings, rendering delays, or session locks can prevent UI elements from becoming accessible to the testing engine.

2. Can I run TestComplete tests headlessly?

Not fully. TestComplete requires an interactive desktop session; solutions include using TestExecute or remote desktop with session keep-alive scripts.

3. How can I detect memory leaks in TestComplete scripts?

Monitor system memory, and handle count during test execution. Also, review script loops and object persistence that may retain memory unnecessarily.

4. What's the best practice for managing licenses in CI pipelines?

Use a floating license server, stagger executions, and automate license monitoring to prevent over-utilization or blocking.

5. Why do tests behave differently between agents?

Differences in OS versions, patch levels, resolution, or TestComplete versions can cause inconsistent test results. Use golden images to standardize agents.

Contact Us