Context:
In order to handle string variable values which are too large for ES, we will add the `ignore_above` setting to the process instance index. This is generally fine for raw object variables values as we only display them in the raw data report, but for flattened object properties, this may lead to slightly inaccurate reports: instances with values that are too large will be displayed as if their variable value was null. We should consider the following options to deal with this:
Option A:
Upon import, replace values that are too long with a placeholder indicating this, eg "Value too large". The values of these variables are not usable for reports/analysis anyway, and this would avoid them being falsely counted as `null` variable values.
Option B:
Consider a truncate token filter: https://www.elastic.co/guide/en/elasticsearch/reference/current/analysis-truncate-tokenfilter.html
AT:
- Evaluate the above options and implement one of them
- Document how large values are handled
Task
L3 - Default