* add async and load_test config for parallel requests to pods
* add async_client.py to Dockerfile
* fix namespace for bacth mode
* take worker exceptions into account and target only one pod in server mode
* do gather after sending all messages
* collect exception in gather