How Microsoft Defender Slowed JetBrains Tools and the Collaborative Fix
Discover how JetBrains and Microsoft optimized startup times by tackling Defender scans.
1The Performance Challenge
JetBrains' ReSharper and Rider faced unexpected slow startup times after transitioning to an out-of-process architecture. User feedback and telemetry revealed Microsoft Defender scans were adding tens of seconds to launch times. This became a critical issue for developers relying on these tools daily, prompting an in-depth investigation into Defender's scanning behavior and its impact on developer productivity.
2Uncovering the Culprit
Profiling identified Microsoft Defender as the source of latency, with scans taking 30+ seconds for JetBrains tools versus <1s for Microsoft IDEs. The key discovery: processes in write-protected paths received preferential treatment. When ReSharper ran as a separate process outside protected directories, Defender performed full scans. This finding was validated through Event Tracing for Windows (ETW) logs across multiple developer tools.
3Measurement Methodology
The team established rigorous testing on a Dell Pro Max 16 laptop with Windows 11 Pro, using Hyper-V VMs to simulate cold starts. They measured scan times across 10 iterations per tool, tracking both ETW logs and CPU time. The data revealed stark contrasts: JetBrains IDEs (10-40s scans) vs Microsoft IDEs (<1s). CLI tools showed faster scans (~2s) due to simpler binary structures.
4Microsoft Collaboration
JetBrains engaged Microsoft's engineering team to understand Defender's scanning principles. Microsoft revealed that files in write-protected folders received optimized scans. This insight led to reviewing installation paths - discovering that Rider's %LOCALAPPDATA% location triggered intensive scans. Microsoft subsequently released Defender optimizations (v1.449.454.0) specifically addressing this case.
5Technical Solution
The fix involved leveraging write-protected directories for installations. When JetBrains tools were installed in these paths, Defender applied lighter scanning. Microsoft's update further optimized this behavior. JetBrains also created a measurement tool using PowerShell's New-MpPerformanceRecording and Get-MpPerformanceReport cmdlets to validate improvements, sharing this methodology publicly.
6Performance Gains
Post-optimization, ReSharper OOP and Rider saw dramatic improvements. Scans that previously took 30-40 seconds reduced to near-negligible times when installed correctly. The chart data showed comparable startup performance to Microsoft's own IDEs, resolving the workflow disruptions developers had experienced. These changes were rolled out in ReSharper 2026.2.1 by default.
7Broader Implications
This case highlights how security tools can unintentionally impact developer experience. It establishes best practices for software installation paths on Windows and demonstrates the value of vendor collaboration. The public measurement tools enable other developers to diagnose similar issues, potentially benefiting the entire Windows developer ecosystem.
8Key Takeaways
The investigation underscores the importance of: 1) Understanding security tool interactions, 2) Vendor partnerships for deep technical solutions, and 3) Robust performance measurement. For developers, it means faster IDE startups; for Microsoft, improved Defender behavior; and for JetBrains, enhanced user satisfaction. The open-sourced tools continue to help others optimize their Windows software.