Hi all,
For some of our elasticubes we use custom code to e.g. use fuzzy string matching. However we noticed our RAM usage for custom code amounts to 6GB.
Can anyone else here give me some pointers on how much RAM usage I should expect?
Is there any article covering best practices for using custom code in Sisense?
Should we explicitly write code statements clearing memory after executing all cells?
Kind regards,
ScheurK
Hi ScheurK,
There is an option to run customcode in the same pod as the bld pod of the EC.
From the CLI:
si datagroups update -name Default -custom-code-mode INTERNAL_CUSTOM_CODE
Assume your data group is Default
This will make sure everything cleanup after the build.
In addition you can use DASK instead of pandas so you can work on a very large data set without being everything in memory. There is example for that.
In any case memory limit of custom code pod can be increased.
Ariel
