The pain is this:
We are running k-gap components on a shared platform and provide access to the jupyter instance to multiple users at the same time.
Because the jupyter instance is by default removing login and thus virtualization per login, we risk having unintended consequences of work by one person influencing that of another.
We need to investigate and document how to best organize for this.
Note: this might be not that easy -- kgap at inception has been conceptualized as a localhost and single user tool. So making that clear and "do anything else at your own risk" might be the safest strategy.
The pain is this: We are running k-gap components on a shared platform and provide access to the jupyter instance to multiple users at the same time.
Because the jupyter instance is by default removing login and thus virtualization per login, we risk having unintended consequences of work by one person influencing that of another.
We need to investigate and document how to best organize for this.
Note: this might be not that easy -- kgap at inception has been conceptualized as a localhost and single user tool. So making that clear and "do anything else at your own risk" might be the safest strategy.