My best results were with standard (in those days) gluteraldehyde fixation, embedding in Epon, and sectioning with diamond knife instead of glass knife.
I never tried embedding in agarose — and it doesn’t make sense to me to do it because you’ve got to be able to chemically crosslink the wall with the embedding material to prevent the bacterial slice from falling out of the thin section.
Use conventional methods such as fixing in Gluteraldehyde followed by dehydration, embedding in resin and eventual sectioning using an ultramicrotome and final viewing and documentation of interested parts.