Yeah, the templating error went away. I’m just trying to validate the cgroup memory usage info. I submitted a stress job running at 96GB of RAM, but the cgroup usage graph only shows ‘4.29B’, so I think there’s something off in the calculation. Prometheus is showing a rising number: cgroup_memory_used_bytes{cgroup="/slurm/uid_1000/job_3367520"} 7.481991168e+10. I’m going to leave the job running for awhile and see what happens with it.