Add debug workload feature

Now we can see what happened at each span in pod lifecycle. It will be great to extend the feature to workload, like replicaset and statefulset. Then we could also get a global view at workload level, which will help us to find which pod is slow or gets error, and why the whole workload is not healthy.