Using flux I need to group over a field, is it possible?


Using flux I need to group over a field ID_Process (its a PID) , is it possible?

|> group(columns: ["ID_Process", "instance"])

(field) Process_ID is a PID for the Process
(tag) instance is tag for Process name

Reason: Line graph must show as one series instead of treating them as two different fields
Expected: Process.last{instance: WmiPrvSE, Process_ID 3968} 13.8
(see grafana snippet)


Due to risk of Cardinality explosion, cannot add/replace PID as a tag (replace the ### in the name)

Tried following script but still it output two rows.

from(bucket: "general_server/autogen")
|> range(start: -1m)
|> filter(fn: (r) => r._measurement == "Process")
|> filter(fn: (r) => r["_field"] == "Percent_Processor_Time" or r["_field"] == "ID_Process")
|> filter(fn: (r) => r["host"] == "")
|> group(columns: ["ID_Process", "instance"])
|> drop(columns: ["_start", "_stop", "CDH", "Env_Class" , "Env_Tags", "Products", "Roles", "_measurement", "objectname", "Environment"])