A host from our cluster has a hardware defect. Indeed, our supplier OVH confirmed to us that the host went off automatically to prevent overheating due to a fault on the water-cooling.
Our applications have automatically restarted on another host of our infrastructure involving downtime.