Your advisor is correct, this is not a plagiarism issue.
Whenever possible, questionnaires should use validated survey questions and responses from previous research. These should be used word-for-word if possible, as small language changes can affect how respondents answer.
If you are totally making up new survey questions when there are good questions that measure constructs out there, you are probably making a mistake.
If you are changing a word or two here and there solely to be “original”, you are probably making a mistake.
Cite appropriately, of course. Usually citing the original source of the validated questions is best, and maybe one or two other influential papers that are directly related to your topic.
Note that you do not need to cite every paper using a similar question to yours. There are fields/questions where hundreds of papers use the exact same question—imagine how long those citation lists would be if you had to cite every other paper with a similar question.