Currently we are using Influxdb on our testing server. We are pushing data from Telegraf( after every 10 sec), Netdata(10 seconds) and some data about vehicles from our django app( after every 1 min).
I have been using it for like 3-4 weeks now. But now, it is causing problems for us. The server has very low RAM( 1675 M ), and influxdb is taking more and more RAM day by day.
It usually takes between 500-800 mb. But today, influxdb crashed and did not start as it goes OUT OF MEMORY and cannot be allocated any memory. The RAM usage risen to 900mb and it was having trouble reading the next WAL files( I checked the log files ), and it could not start. We freed some RAM, and it started , and is taking around 1200 mb of RAM.
This question is important for architecture of our next server which we are going to use in production, so please direct me in the right way. We will be using influxdb for the same processes as defined above.
So, my question is that Does the RAM usage of influxdb will keep increasing if I keep pushing more and more data day by day ? It will increase for sure, but any idea of the maximum amount of RAM which can be occcupied by Influxdb if I have to push and store data for 2, 3 years ?
This is the /var/lib/influxdb directory’s size currently,
alphauser@AlphaServer:/var/lib/influxdb$ sudo du -hsc * 329M data 8.0K meta 91M wal 419M total