The execution in matched pods with low efficiency

[StephenLiuYa]

Issue Description

I want to let thousands of pods loss network in time,but I found the execution in matched pod is serial，so it is inefficient

• feature request

Describe what happened (or what feature you want)

• test case

apiVersion: chaosblade.io/v1alpha1
kind: ChaosBlade
metadata:
  name: loss-pod-by-labels
spec:
  experiments:
  - scope: pod
    target: network
    action: loss
    desc: "node network loss"
    matchers:
    - name: labels
      value: ["app=nginx"]
    - name: namespace
      value: ["default"]
    - name: percent
      value: ["100"]
    - name: interface
      value: ["eth0"]
    - name: local-port
      value: ["80"]

• result:

time="2021-05-14T13:58:20Z" level=info msg="get pods by labels app=nginx, len is 51" experiment=loss-pod-by-labels
……
time="2021-05-14T14:08:22Z" level=info msg="trigger update event, name: loss-pod-by-labels"

It takes 10 minutes to loss network in 51 pods