Self-pressurized rapid freezing (SPRF) has been proposed as a simple alternative to traditional high pressure freezing (HPF) protocols for vitrification of biological samples in electron microscopy and cryopreservation applications. Both methods exploit the circumstance that the melting point of ice reaches a minimum when subjected to pressure of around 210 [MPa], however, in SPRF its precise quantity depends on sample properties and hence, is generally unknown. In particular, cryoprotective agents (CPAs) are expected to be a factor; though eschewed by many SPRF experiments, vitrification of larger samples notably cannot be envisaged without them. Thus, in this study, we address the question of how CPA concentration affects pressure inside sealed capillaries, and how to design SPRF experiments accordingly. By embedding a fiber-optic probe in samples and performing Raman spectroscopy after freezing, we first present a direct assessment of pressure buildup during SPRF, enabled by the large pressure sensitivity of the Raman shift of hexagonal ice. Choosing dimethyl sulfoxide (DMSO) as a model CPA, this approach allows us to demonstrate that average pressure drops to zero when DMSO concentrations of 15 wt % are exceeded. Since a trade-off between pressure and DMSO concentration represents an impasse with regards to vitrification of larger samples, we introduce a sample architecture with two chambers, separated by a partition that allows for equilibration of pressure but not DMSO concentrations. We show that pressure and concentration in the fiber-facing chamber can be tuned independently, and present differential scanning calorimetry (DSC) data supporting the improved vitrification performance of two-chamber designs.