It seems there have been some issues with Kubernetes jobs as infrastructure blocks, potentially related to the message in your log. Here are some key points from the GitHub issues that might relate to your situation:
1.
Labels not being applied to jobs on a kubernetes work-pool (
#11172):
- A user reported they're unable to apply "azure.workload.identity/use=true" labels on jobs created from a Kubernetes work pool.
- They tried to set labels through the UI and directly in the job manifest, but jobs still lacked the expected label.
- This seems to be a bug with the version 2.14.2 of Prefect.
2.
Ability to customize behaviour of Kubernetes Client (
#8572):
- Another issue mentioned difficulties running k8s jobs on different clusters.
- The user faced problems due to SSL certificate verification failures between the Kubernetes client and the API server.
3.
Prefect 2.7.0 (cloud): Failed to get infrastructure for flow run (
#7796):
- A user experienced intermittent failures where flow runs got stuck in a pending state, possibly due to issues with the Kubernetes Job infrastructure.
Your issue could be related to one of these reported problems. It may be helpful to compare the error messages and behaviors noted in these issues with what you see in your log to determine if you are experiencing a similar problem. If these issues don't seem to match your situation, additional details from the logs you are referring to would be necessary to further diagnose the problem.