Hello Influx
I have been using Influx successfully for ~2 years now and recently upgraded to 1.5
I collect metrics from 100s of Nodes running model (ML) training jobs. Metrics are around Hardware (steady stream irrespective of load) + metrics arounds jobs (burst or steady or low).
As i run open source version, i have shard-ded the metrics into 5 categories with single instance for each, Each influx instance is running as a Kubernetes POD with ~200Gigs of RAM and multiple cores.
I now want to measure total number of points ( I do ingestion with telegraf and batch mode) per influx instance per second.
I am able to obtain number of successful HTTP POST per second.
$ cat h.log | grep POST | grep "31/Oct/2018:14:49:10" | grep 204 | awk '{print $10}' | sort | uniq -c
1 200
170 204
$
But that does not tell me, number of points written per second into Influx. As each HTTP POST contains hundreds of points (batch + buffer).
How do I measure, number of points ingested per second ?
Regards,
Deepak