- I have an in-memory distributed storage system which contains more than 100 instances
- I would like to monitor the overall memory usage
- Telegraf will write the metrics to InfluxDB every 10s, but there may have sometimes Telegraf writes to InfluxDB timeout (very few and acceptable).
The proposed query looks like:
SELECT SUM(MEAN(used)) FROM “telegraf”.“autogen”.“mem” WHERE $timeFilter GROUP BY time($interval) fill(null)
But this is a syntax error because SUM needs to have a field parameter.