Context:
To provide autocompletion in the string variable modal and also allow to filter for variable contains/not contains, we use a wildcard query to filter for right variable values (if the given value is larger than 10). However, this has several downsides:

• the ngram analyzer, which is used for variables with values less than 10 signs, will explode for large variable values since it creates all possible combinations of 10 signs
• the wildcard query has a prefix wildcard (e.g. *myValue), which is not recommended by Elastic since it slows down the query significantly.

Instead, we should:

• decrease the value of the ngram analyzer to 5 to mitigate the explosion problem.
• have an additional multivalue field which is the reversed version of the current value. Then you can do two a postfix wildcard query (e.g. eulaVym*) for the initial prefix query (e.g. *myValue)

Hint:

• We might need to think about how to reduce the ngram value in the upgrade script