Without looking into this specific outage, I’d suggest things like deferred maintenance and “cost optimizing” technocal staffing are often contributing factors. (At least in my experience)