Way more contagious variant that spreads rapidly among people w/o prior immunity? At least that's what is being theorized based on what happened in the UK (very sharp rise in cases, followed by pretty sharp drop-off, although that seemed to stabilize now).
Vaccines have several effects. Frequently they prevent you from getting serious symptoms (or any symptoms at all) after contact with the virus. Often they reduce the time span during which you are infectious, sometimes they mean you aren’t infectious at all after contact with the virus. Sometimes (very rarely) they have no effect at all.
Only that last point can reasonably be framed as the vaccine not being perfect.
Other than that the vaccines work astonishingly well, even with delta. Them still making it possible for you to be infectious after contact with the virus doesn’t really change that.
Switching to proportion is actually the wrong thing to do. If you look at the data from a place with high vaccination rates (such as King County), even low probability of a break through case among the vaccinated will translate to high proportion.
Edge case that demonstrates the fallacy of looking at the proportion is: let's say you have 100 people, 90 of them vaccinated, 10 not vaccinated. Let's then say that vaccinated have 10% of getting infected, and the unvaccinated have 50% of getting infected. You'll then get 9 vaccinated people that got sick, + 5 unvaccinated people that got sick, for an almost 66% proportion! But that doesn't change the fact that as a vaccinated person, your chances of getting infected are 5x smaller compared to the unvaccinated.
As for the effectiveness of vaccinations, I'm optimistic based on the data from King County, WA: https://kingcounty.gov/depts/health/covid-19/data/vaccinatio...