Due to cooling issues on the 2nd and 7th of January there where short outages in the HPC2N region and all the running instances where shut down unexpectedly. The underlying issue causing these cooling issues has been resolved but you might need to start up your instances in the cloud again manually.