Slurm integration

With Slurm integration active, you can easily see the jobs running on your system, search through job history, and look at detailed records for each job.

Integration with statistics

The job details page shows every ClusterVisor statistic gathered from the nodes running the job, for the job’s runtime. So you can correlate CPU usage, GPU usage, memory usage, CPU temperature, network I/O and any other stat ClusterVisor collects with your users’ Slurm jobs.

Correlating node statistics with job details helps administrators answer important questions like how much memory users need for their jobs, and helps diagnose support requests like “my job is running slowly.”

How it works

• Works with a recent version of Slurm and the Slurm REST daemon
• No need to set up processes to ingest job data into ClusterVisor
• No data duplication: running jobs and job history are queried directly from Slurm via the REST daemon
• Full query ability for all jobs, even ones from before ClusterVisor was running