[Spot] An option for keeping failed spot job for a while before termination

[Michaelvll]

When something wrong happens with the spot job, it would be nice to be able to log into the spot cluster to take a look at the problem. As proposed by @lhqing, having an option like --keep-minutes-after-error 60 for the spot launch can be useful for debugging.

[github-actions]

This issue is stale because it has been open 120 days with no activity. Remove stale label or comment or this will be closed in 10 days.

[Michaelvll]

This was requested by another user as well. We should keep this issue open.

[github-actions]

This issue is stale because it has been open 120 days with no activity. Remove stale label or comment or this will be closed in 10 days.

[Michaelvll]

A new user asked for this. We should consider adding this.

This is mostly a workaround for the long startup times; If I make some sort of mistake, I would like to be able to ssh into the vm and run some new commands without needing to wait through the startup time again