Open JackGzj opened 6 years ago
It's one of the things I planned to explore, but didn't get far enough. It's fairly easy to simulate network partitions, but the simulation isn't really simulating latency in a useful way, it's more of a topology simulator.
We have been using the spigo for a while and our team is trying to simulate the latency faults just like the working mode of chaosmonkey, we define the delayvictim in the arch file and when the delay is triggered, a message will be sent to the node which will take the operation of sleep. We are still correcting the effort so that the corresponding tracing data can be labeled. And there is a new question: what does each "package" configuration means in the arch file? I am defining my own arch but it doesn't work as expected.
Is there some ways to config or inject some faults( like latency added ) to some service? It's quite important to simulate the cluster running situation.