Why is the kernel couldn't be connected to the IBM Watson studio platform?
Sometimes the kernel cannot connect because of issues in the underlying environment. If you've tried to modify the python version, Jupyter version, etc through your conda yml file, you could cause issues with the environment.
You can also face issues with the kernel connecting if you kill the kernel based on some computation like you try to invert a very large matrix but you don't have enough memory. In that case, try adding more RAM.