The RQNs are not always reliable. The instrument software uses an algorithm to determine the RQNs and it is easily fooled. The algorithm method is proprietary, but we do know that false positives can happen when the program determines part of your sample as a rRNA peak that is not actually a rRNA peak. False negatives can happen for the same reason, except that it misses the actual rRNA peak.
This is because the algorithm is based on mammalian (eukaryotic) RNA. If your RNA is not mammalian/eukaryotic, the software may not detect the actual rRNA correctly, and might miscalculate the RQN. The RQN can be a helpful guide at times, but your visual interpretation of the data is better.

Below is a helpful guide from the Bioanalyzer. It refers to the RQN as RIN, but they are interchangeable.
RIN Document

A helpful paper on insect RNA quality.
https://pubmed.ncbi.nlm.nih.gov/21067419/