This evening the temperature controller of the cristal1 of the slave laser shut down (not really surprising).
I have switched it with a spare one. Everything seems back to the standards conditions.