Pick one and press Start, then click Next › to step through it.
Watch for
S4's in-flight count climbing wave after wave while the fast servers finish and go idle — equal by count, unequal by load.
More about this scenario
You'll learn
Why counting requests isn't the same as balancing load — an even split by count overloads a slow backend.
How it runs
Round-robin just cycles through the servers in order, giving each the same number of requests regardless of how busy it is. The slow backend (S4) can't drain its queue between waves, so its in-flight work piles up.