[slurm-users] Efficiency of the profile influxdb plugin for graphing live job stats

[Andrew Elwell]

Just realised after I sent message
 > Measurement: acct_gather_profile_task   Tags: job, step, task, host
  Fields: CPUTime, CPUUtilization, CPUFrequency, RSS, VMSize, Pages,
ReadMB, WriteMB    Timestamp

It'd also be useful to (if available) capture "cluster" as a tag too,
- we have three clusters that have the same nodename format (nidNNNNN)
so need some other identifier to narrow down a job, else I'd need a DB
per cluster (also possible, but then I can't select tag values from
cluster as a variable in grafana)


Andrew