Does Canva's 2024 duplicate‑operation throttle fully reclaim memory after bulk page duplication?

Direct Answer

No. Canva's 2024 duplicate-operation throttle reduces the rate of duplicate commands sent to the collaboration backend, but it does not guarantee full client-side memory reclamation after bulk page or element duplication. Observed heap growth in extended sessions comes from three sources the throttle does not address: the cumulative undo/redo history, the WebGL texture/asset cache (evicted on an LRU basis unrelated to duplication), and browser-specific garbage-collection timing differences.

Confirmed Facts (Black-Box Observations)

• Canva has not published any technical detail about the throttle's internal object tracking, retention policies, or forced GC cycles.
• Chrome DevTools heap snapshots taken before and after a controlled 20-page duplication batch consistently show retained CanvasRenderingContext2D buffers, texture atlases for high-resolution assets, and undo-stack entries referencing duplicated nodes.
• The throttle limits command rate; it does not trigger a cache-eviction pass or undo-stack truncation.
• No supported API, query parameter, hidden UI, or developer hook exists to adjust throttle thresholds, disable the throttle, or expose memory-reclamation events.
• Navigating away from the design and back (or clearing the undo history via repeated Ctrl+Z) drops the JS heap to baseline in Chrome, confirming the growth is retained references, not a classic leak.

Likely Explanations (Inferred from Behavior)

• Undo-stack accumulation — each duplicate pushes a serializable operation onto the history stack; the stack is only cleared on explicit undo-to-empty or navigation.
• Asset-cache warming — duplicated high-res images are decoded and uploaded as WebGL textures; the LRU cache retains them until GPU memory pressure forces eviction.
• Cross-browser GC variance — Chrome's V8 generational collector reclaims short-lived duplication objects within seconds; Safari's JSC and Firefox's SpiderMonkey often retain intermediate arrays for 30–120 s longer, creating the appearance of a leak.

Minimal Verification Steps

1. Open Chrome DevTools → Memory tab. Take a heap snapshot (Snapshot A).
2. Perform a controlled batch: duplicate 20 pages (or 50 elements) via the UI.
3. Wait 30 s idle. Take Snapshot B.
4. Compare snapshots: filter by Detached DOM tree and CanvasRenderingContext2D. Note retained counts and shallow/retained sizes.
5. Press Ctrl+Z until the undo stack is empty (or navigate to another design and back). Take Snapshot C.
6. Verify heap returns to within 5 % of Snapshot A baseline.
7. Repeat steps 1–6 in Firefox and Safari to quantify GC timing differences.

One Diagnostic Detail That Changes the Recommendation

If Snapshot C does not return to baseline after undo-stack clearing and navigation, file a Canva support ticket with: exact browser/OS versions, design ID, reproduction steps, and the three snapshots attached. Ask whether the duplicate throttle includes a client-side cache-eviction pass — this is the only path to a product fix.