Updating single trace of many: changing `revision` prop does not improve performance (compared to redrawing all)

Hi, I am trying to redraw only single trace in a plot of many traces (biophysical data). I tried to follow instructions in react-plotly and plotlyjs documentation. Plot is updated but I don't see any performance improvement if I change only single element of data or all data elements, if I change revision as described in following:
https://github.com/plotly/react-plotly.js/#refreshing-the-plot
https://plotly.com/javascript/plotlyjs-function-reference/#plotlyreact
Important Note: In order to use this method to plot new items in arrays under data such as x or marker.color etc, these items must either have been added immutably (i.e. the identity of the parent array must have changed) or the value of layout.datarevision must have changed.

I generate sample data 100 traces of 10000 points with:

  const data = useRef([...Array(100).keys()].map( c=> (
    { x: [...Array(10000).keys()],
      y: [...Array(10000).keys()].map( c=> c/10000.0 + Math.random()/5)
    } ) ))
  const revision = useRef(0)

Also, I change data with:

revision.current++
data.current[0].x = [...Array(100).keys()].map( v => v*100)
data.current[0].y = [...Array(100).keys()].map( c=> Math.random())
setState({...state, data:data.current, revision:revision.current})

Plot is generated using data.current for it's data prop

<Plot
data={state.data}
revision={state.revision}
...

Drawing plot containing single trace takes just a split of second, but if I had already drawn many traces and I want to redraw only single trace, it takes same time as to redraw all traces (few seconds).
Am I doing something wrong?

Start by reproducing the reported case with the 100-trace sample and compare changing one trace with changing all traces. Inspect the Plot component's data and revision update path alongside the referenced react-plotly.js and Plotly.js documentation; done requires establishing whether single-trace redraw is supported and documenting or fixing the observed behavior.