my influxdb was running on a 4 cores 16 Gb VM and we upgraded to a 8 cores 32 Gb VM as we were running out of memory. In just a few days, we are already consuming 30 Gb.
The memory consumption seems to be synchronized with the write load, not really with the query load.
I have extracted a few lines from the show stats output…
pointReq pointReqLocal req subWriteDrop subWriteOk writeDrop writeError writeOk writeTimeout
8876222 8876222 1671558 0 1671558 0 0 1671557 0
How can I control the memory consumed by Influx?
I don’t need to keep several weeks of data, so if I start to export old data and drop the oldest shards, will it help the memory consumed by Influx.