Lacks a way to record that a CI run was terminated
Sometimes a CI run never ends, and is terminated due to external actions, such when the machine running CI is rebooted without warning. The CI broker has no way of noticing this, and worse, no way of recording that this has happened.
If a CI run is terminated due to taking too long, it is marked as failed. Runs that have started and never finished could be marked as failures as well, either when the CI broker starts, or via
cibtool.cibtool run updatecan already do it one by one, but we could add a subcommandcibtool run fail-runningor something.