Playing around with a ICM7555 variety of the 555 timer, datasheet advertises delays from us to hours. This version has much lower input currents so in theory, larger resistors can be used compared to the original 555's.
If you use a large value resistor > 1M

and a large "leaky" electrolytic capacitor I suppose you might get something to work beyond say 60 minutes. I would guess that the timing would vary a fair bit over time and temperature though because the capacitor leakage current would have a significant impact on the timing.
The datasheet shows R value as high as 100M, I'm not sure how practical that is for real world use with eventual contamination on the board etc? I did a reality check with a 4.7M

and, 47uF Tantalum and that yields about 16 minutes in astable mode, however with a 10M

and the same capacitor it never timed out.
Assuming using a divider is not an option, has anyone been able to generate long time periods and or really low frequencies without an electrolytic?