Rasa/duckling - ready for production and scaling?

hi all,

is the image " rasa/duckling" ready for production?

  • how many parallel requests does ONE instance of “rasa/duckling” can handle?

thx, r