as mentioned in the title i am asking for a best practise advise, if i should install the latest version of a library or module (e.g. for Python), when building an Production Image via Dockerfile or specific version or should i do it total differently?
Right now i use install commands like so to get python modules:
RUN pip3 install selenium
RUN pip3 install bs4
RUN pip3 install pandas
in my Dockerfile.
My doubts are based on the fear, that my Production Image could be different form the one on my Testing-stage. To complete my scenario: i use Cloudrun on GCP to build my images.
Pls provide your best practise advise, how to deal with (python)-modules when building an image for Production.