We are currently undertaking the process of reimaging and upgrading Ubuntu in our FoundationDB cluster. To ensure a smooth transition, we are performing this activity on one node at a time while keeping the node in maintenance mode. After a node successfully joins the cluster and the cluster returns to a healthy state, we proceed with the upgrade on the remaining nodes.
During maintenance activities on the coordinators in our cluster, we have observed that our monitoring app intermittently reports multiple nodes as being down for approximately one minute before they come back online.
why this is happening only while performing activity on coordinators?