When we add a node to a cluster or remove a node from a cluster, data redistribution will be involved. When the disk failure of a node recovers, the data of this node will be backfilled. So when these happen, how do we control the impact of traffic brought by these operations on business traffic?
Will scylladb have a business traffic priority mechanism?
ScyllaDB already has CPU and IO schedulers. User workloads and background maintenance work, like repair, streaming and compaction use different scheduling groups. These are isolated from each other and have different priorities: the user workload has 1000 shares and the maintenance workloads have only 200 shares.
Furthermore, ScyllaDB Enterprise has Workload Prioritization, which allows for isolating different kind of user workloads from each other, assigning them different priorities (shares).
1 Like