Trainspotter, you are correct that the last swath on July 17 was at 9:55 UTC (more than 3 hours before the MH17 crash) by Aqua, and the first swath on July 18 is by Terra, at 8:55 UTC.
Still, since the swaths of Aqua and Terra overlap 50% in the Ukraine area, there are TWO images for each day.
Thus, we now have 2 images for July 17 and 2 for July 18, as you can see by selecting the data and the satellite in the WordView link :
July 17, 8:12 UTC, Terra : mostly clear skies above all of East Ukraine, scattered clouds in Donesk area.
July 17, 9:55 UTC, Aqua : (screenshot 1 above) mixed level clouds around Donesk, Lugansk scattered clouds.
July 18, 8:55 UTC, Terra : (screenshot 2 above) solid cloud cover all around east Ukraine.
July 18, 10:40 UTC, Aqua : solid cloud cover all around Ukraine.
I understand your point that in order to find cloud cover during daybreak on July 18, we should not extrapolate (interpolate actually) cloud cover the pictures from July 17, 9:55 UTC, Aqua and July 18, 8:55 UTC, Terra.
But there are other satellite pictures around, so we do not have to interpolate.
For example, there is MeteoSat, in geostationary position, which takes visible light and IR images which are released here an hourly basis :
http://www.sat24.com/history.aspx
East Ukraine is in on the far right in these images.
Set the time and date, and notice that during the morning of July 17, Eastern Ukraine was mostly cloud free, with clouds moving in from the South West during the day, which turned to solid cloud cover as July 18th started (even in IR; which means these clouds are thick), and even got thicker during the course of July 18th, only to slowly dissipate during the 19th.
This shows that the morning of the 18th, Lugansk has solid overcast, while the morning of the 17th Lugansk had clear skies. Does that make sense ?