fluid
fluid copied to clipboard
[ENHANCEMENT] High load on CPU and memory when using Fluid in a large-scale cluster
What enhancement you'd like to add: When using Fluid in a large-scale cluster where there is a large amount of pods in one single namespace, JindoRuntime controller requires very high CPU and memory request. For example, if 20k pods and 20k datasets are created in the same namespace, binding 20k datasets require about > 8 CPU and 16GiB Memory.
Why is this enhancement needed: