Test performance of diff collection and single diff limits
Everyone can contribute. Help move this issue forward while earning points, leveling up and collecting rewards.
From #254823 (comment 419025968):
I would love to see whether we could take a step back and shift the conversation here away from file size to impact. Could we measure what 10kb actually means for the loading time on different network speeds, and what the total number and size of all files means for responsiveness and lag on different machine types. If we have that data, we could make a more informed decision what the actual limit should be.
Our current diff limits seem to have been arbitrarily picked. How can we make a more informed decision on these important values?
For example, what is the performance difference of a 5 MB collection of file diffs on a high-speed network and a performant device vs. a low-speed network and not-so-performant device?
What is the average network speed and performance of devices of GitLab.com users?