chore(workflows): enable fail fast strategy

[axel7083]

Description

This may help reducing the number of jobs running on the organisation

[benoitf]

I know that during a call, there were a request from maintainers to know all the failures and not the first one b/c at some point it was fail fast

(for example: did I just forget the linter or it's the linter and unit tests and also e2e)

[axel7083]

(for example: did I just forget the linter or it's the linter and unit tests and also e2e)

I understand the remark, but this is a comfort that is costly 🤔 chaining and fast failing reduce a bit the consumption when the developer commit something that has an issue.

This model could be chaining further, putting the linting as a condition for all other checks. More in depth analsys finding out which check is failing the most would be interesting to properly organise the CI to reduce usage.

For example, I would not be surprised to only start e2e if unit tests are passing, e2e are very costly and long task, meaning they probably should only run to ensure no regression