How do we explain the drops in requests in the following example:
Shouldn't it always be the same load ?
Do we know if it's an it's an issue in APM, or is it that the load
balancing "lose" some requests as we scale down a server ? or else ?