Issue Description

Type: feature request

Describe what happened (or what feature you want)

chaosblade是一款覆盖面很广的故障注入测试工具，因此我希望引入此软件对 K8s Paas平台进行测试，但是发现只能针对特定的pod进行故障注入（delete pod或者pod cpu负载等），没有其他针对Paas 的chaos工具（Powerfulseal、Chaos-mesh等）拥有在规定时间内随机等delete 规定分区下的po

There is no e2e test and documents for external network delay and partition now 🤦 .

The e2e test could be found in /test/e2e/chaos/basic.go. Reading other tests in that file could help accomplish this issue.

Example for extern network delay:

apiVersion: chaos-mesh.org/v1alpha1
kind: NetworkChaos
metadata:
  name: network-netem-example
  namespace: chaos-testing
spec:

This issue is

One problem with faulting programs under KRF is that KRF might decide to inject a fault during the dynamic link/load phase, aborting ld-linux.so instead of the actual target image. This usually isn't helpful, since it doesn't indicate any mistakes in the target itself.

It should be possible to check the loaded program's name via the current task, probably via comm. We should use that (or