Hello! Is there a way to predict influxdb disk usage for a measurement or set of measurements if we have an estimate of the number of incoming points ? I tried measuring it using _internal
database and also the size of the docker volume on disk . However I’ve noticed that it doesnt constantly keep increasing by the hour . It dips suddenly and rises again slowly . So it becomes a bit difficult to predict how the disk size would grow over a period of time
Hello @prashanthjbabu,
Forecasting disk usage is a complex problem for InfluxDB as it depends on the compaction planner, retention policies, shards, and more. You could try using a forecasting algorithm.
You might want to investigate:
- ARIMA
- Holt Winters
- FB Prophet
- LSTM
Here is a repo with some examples that might be of value to you GitHub - InfluxCommunity/Notebooks: A collection of Jupyter Notebook tutorials on anomaly detection, forecasting, and InfluxDB.
Thanks @Anaisdg . will take a look at this !