I'm heard here, that ceramic metal halide lamps, have lower internal pressure than quartz metal halide lamps, so that violent EOL is less frequent. Despite this, they takes much longer time to cool down before restriking compared to quartz metal halide.
Why?
The answer is not that complex, more accurately it could be described in two points:
1) The arctube operates at a higher temperature. That means it obviously takes longet for the arctube to cool down to a state, when there is just the buffer fill in the gaseous form. As that means condensation of the majority of the mercury fill, the restrike temperature is about the same as with quartz counterparts. So with the difference being larger, it means there is more heat "stored" in the mass of the arctube, so it takes longer for it to go away.
2) As the operating temperature is higher and thermal losses lower (higher efficiency in converting the electricity to all radiation means there is less power available to heat up the arctube), the complete assembly means thermal resistance from the arctube to the environment should be made higher (practically done utilizing hard vacuum outer instead of the gas fill in the QMH's). While the mass of the arctube is about the same as the quartz counterparts (the burner dimensions of the modern quartz lamps are not much different), the higher thermal resistance means the arctube cools down slower. Plus the heat transfer is mainly by radiation, which gets way slower as the arctube is cooling down, making the cooling down even more slower.