[Feature] Performance problems when the cluster is running on a small hashrate

[renajin]

Feature Request

Is your feature request related to a problem? Please describe: If a transaction peak come, how to scale-out immediately without affecting current sql process, or can we have any other method to handle this emergencies?

Describe the feature you'd like:

Describe alternatives you've considered:

Teachability, Documentation, Adoption, Migration Strategy:

[renajin]

When initially creating small size cluster, we set cpu limit to 1/4 of the maximum hashrate, so this cluster can withstand a certain amount of transaction peak.