These are chat archives for influxdata/influxdb

Dec 2016
José Eduardo Quadrado Correia
Dec 12 2016 15:22
Hello all! I don't have a problem per-se but I'm running into something that maybe is some noob-ness on my side. At my organization we are using InfluxDB (right now the 1.0) for our monitoring metrics for over an year now without complaints. Since we are in the IoT sensor business we came up with the scenario of using InfluxDB to store our sensor data and use it as our main DB (not the current scenario at the moment).
I've setup a simple testing scenario with a single node (8 core, 16GB RAM box) and inserted 15M rows into a single measurement with 2 tags and 3 fields which models perfectly our data needs.
Using chunked queries I've reached a read speed of about 1M rows every 15 seconds. Changing the size of the chunks alters the behaviour but doesn't really do anything speed wise.
It's this an expected read speed for this kind of scenario? I'm using the autogen retention policy no configuration tuning whatsoever.
José Eduardo Quadrado Correia
Dec 12 2016 15:27
For these tests I'm using the 1.1 release