Skip to content

Analyzing Sensor Storage Performance

Mat Davis edited this page Nov 27, 2019 · 1 revision

After each successful sensor run, the results of the run are stored in the database during the storage phase of that sensor. If database performance (or storage server performance) is sluggish these storage times can become long. It can be helpful to know which sensors are taking the longest to store their results, because then you can analyze whether that sensor is needed or if there are configuration changes you can make to that sensor to help performance.

The attached script can be used to analyze a discovery run and give total and average storage times for all the sensor types that ran and stored results. Download storage.jy and additionally for Windows storage.bat to $COLLATION_HOME/custom (create if necessary). Run storage.bat (Windows) or storage.jy (Linux) without any arguments to see the usage. Output is in CSV format with a header describing the columns.

Note that the script assumes that the discovery was run to completion. If the discovery was aborted or if sensors end in a storage error this might affect the results of script output.

Clone this wiki locally