In short, these validators may begin to lose potential emissions XRD or be excluded from participating in consensus for some time. This is important to ensure that node-runners have a strong incentive to keep the network stable, reliable, and performant.
Of course, in the real world, things happen. Regions go offline; cables get cut; data centers get flooded. Unfortunately, the Radix Protocol can not establish a “reason” for a validator node not participating in consensus. We recommended that validator node-runners employ strategies to ensure high availability, such as those presented in our technical documentation.
The good news is, as long as such problems are infrequent, validator node-runners and stakers won’t miss out on many emissions. Penalties are calculated per-epoch, and with epochs occurring roughly between 30 - 90 minutes, a node-runner that brings their node back online quickly will also quickly pick back up earning rewards.
Further reading: