Why do I get very large (100 percent) confidence intervals for predicted probabilities from logistic regression

Question

I performed this logistic model:

Model2i <- glm(Child ~ Age +Age2 + Education + `Place of residence` +`Marital status` + `Multiple birth`+ Region    + Period*`Conflict exposure` +`Sexual violence`+ `Child's death during war` * Period+`Sibling death during war`+ `Proportion of children died before 1996`, data=DATA2007_10km, weights  = Poids, family=binomial("logit")); summary(Model2i)

I want to plot predicted probabilities using ggeffects as follows:

ggpredict(Model2i, ci_level = 0.95,terms = c("Period","Conflict exposure")) %>% plot(connect.lines = TRUE)

but I get very large intervals for predicted probabilities as seen on the following picture:

enter image description here

I was expecting narrower prediction intervals. Can anyone tell me the cause of the problem and how to fix it?