Hello!
Occasionally our FDB clusters enter a recovery state, which I understand is normal. During recovery, when data is being moved or rebalanced, we regularly experience total saturation of our disk throughput.
Aside from sizing up our disk throughput, are there any knobs or configuration that allow us to limit the rate at which data is moved/migrated? When our disks become totally saturated, it slows down the cluster recovery process and can sometimes trigger more recoveries.