However, batteries drained as low as 2.5 should have minimal degradation of the battery chemistry.
@KeroZen what are your thoughts?
The problem when the open circuit voltage reads 2.5V is that it means that under load the cell went beyond that figure and into the red zone. Every time you go there it creates complex physical faults inside the substrate, like micro tunnels and funky shit... and in the end it damages the cell by raising its internal resistance.
When I push mine they end around 3.3V (open circuit that is) but we're talking about cells rated to go down to 2.5V. There are indeed some cells (mostly older gens and non-IMR/INR chemistries) that should not go below 3V, and super high discharge Li-Po packs for instance are even higher and should not go below 3.3V. But in our case I think all recommended cells are good down to 2.5V.
Anyways and thankfully the device is self-limiting, when you go below 3.5V it starts to become pushy. If you have to double your trigger time, it's a clear sign that you need to swap. The problem is that if you insist you will still get vapor. I once ended below 2.7V this way and I was not happy and the cell was becoming hot too: not good!
As a rule of thumb now, when I start a session if the cell is already below 3.6V I swap it. If it's at or slightly above 3.6V I know I will get a pushy session but I do it anyways and it ends around 3.2-3.4V. It's perfectly fine to swap earlier but you need to keep in mind that these cells are rated for a certain number of cycles, and even if you under-use them it still counts as a cycle (well ok it's slightly more complicated than that, you'll get more cycles it's true but there's still a correlation here)