@Ricardo Jorge de Sousa da Fonseca I think Azure Machine Learning builds the final Docker image from this environment specification when the environment is used in a job or deployment. So, while creating the job the compute used can be of higher capacity to build the environment sooner if possible or you can configure the job with a higher timeout if an option is available like below:
If this answers your query, do click Accept Answer
and Yes
for was this answer helpful. And, if you have any further query do let us know.