We find that amplicon sequencing is very sensitive to concentration and low concentrations output either a low confidence assembly (that is shorter than expected because of a truncation at the 5' or 3' ends) or nothing at all. Even with a high quality assembly you can expect to "lose" the first 10-15 bp with this method.
Our library preparation method shears the amplicon and so you can expect to see a range of different sized reads on your _length.png plot.
The most common failure modes for nanopore amplicon sequencing is too low sample concentration. Although other providers recommend sending Qubit quantified DNA we find that this can be quite the ask for most groups. We ask that you run an electrophoresis gel when you first send us samples to check the quality of your amplicon preparations and ensure the spectrophotometrically measured DNA concentrations make sense given what you see on your gel. Concentrations may require some tuning, but once you find your "sweet spot" you should have very few issues in the future.