[cluster-trace-gpu-v2023] Why pods that're still running have deletion_time?

[AlpacaMax]

In cluster-trace-gpu-v2023 dataset all the pods that are still in the Running phase have a deletion time. Why would they have a deletion time? My understanding is if the pod is still running then the pod hasn't been deleted yet. And if the pod already finishes running its task then the pod_phase should be Succeeded or Failed. Please let me know if my understanding is correct and explain to me how to interpret deletion_time for pods in the Running phase.