A cellar node restarted, timeouts or HTTP 500 errors sent

A node from the old Cellar cluster restarted at 21:30 UTC. While it went okay at first thanks to the restart of the few nodes a few days ago, it started emitting HTTP 500 errors or timeouts, as it was before. Service should be back online in a few hours once the cluster stabilized itself again.

The new Cellar cluster is not impacted by those issues.

EDIT 23:40 UTC: Cluster now seems to be in a good shape again