SwiftLee's SwiftUI Performance: Find Redraws in Xcode Instruments
Original post SwiftUI Performance: Find Redraws in Xcode Instruments
Article Summary
Antoine van der Lee reveals how SwiftUI performance problems hide in plain sight. Most developers know views redraw too often, but measuring the actual cost requires knowing exactly where to look in Instruments.
This deep dive from SwiftLee walks through the complete workflow for finding and fixing SwiftUI redraw issues using Xcode Instruments. It covers the SwiftUI template, interpreting the View Body Updates track, and tracing redraws back to their root causes.
Key Takeaways
- Use SwiftUI template with Release builds to see real performance data
- View Body Updates table shows frequency plus duration per redraw
- Root Causes column reveals which property changes trigger each update
- @Observable macro prevents updates from unread properties
- RocketTrace automates detection of seven SwiftUI performance patterns
Finding expensive SwiftUI redraws requires combining update frequency with duration metrics in Instruments, then tracing root causes back to the triggering dependencies.
Continue reading
Getir
Aug 12, 2026
More from SwiftLee
MVVM in SwiftUI: Using view models without overengineering
SwiftLee
Aug 31, 2026
Also on iOS
Xcode 27 replaced Simulator.app. Expo CLI and expo-device-hub already know.
Expo
Oct 20, 2026
What This Article Covers
Antoine van der Lee ran the RocketTrace Slow Lab workload, which published 600 state changes at roughly 60 Hz. This triggered a 27-second recording with 228 long updates, totaling 1.10 seconds.
He picked a timeline range and looked through the View Body Updates hierarchy. WorkloadCard cost 8.35 ms and InvalidationWorkloadCard cost 6.71 ms, out of 24.53 ms total.
The table showed InvalidationWorkloadCard and WorkloadCard bodies firing repeatedly, each only 40-50 microseconds. That cumulative main-thread cost came from these frequent, cheap updates adding up.
This page summarizes “SwiftUI Performance: Find Redraws in Xcode Instruments” by SwiftLee, published Oct 19, 2026. All credit for the original work goes to the authors.
