[Question/Enhancement] about /status handler

[dene14]

As far as I've understood /status handler on every node returns cluster state and map from its point of view. It also has just 3 statuses (from what I've seen during my tests): NORMAL, DEGRADED, STARTING, which are cluster statuses basically.

For a purpose of self-healing we need some node-related stats to understand if it's degraded or impaired. For e.g. as #1689 states, nodes should be restarted after coordinator restart. If there is at least an evidence that THIS node has problems contacting coordinator node it could be possible to handle with /status health check. Not a solution for a real clustering bug but at least a temporary workaround that should be also easier to implement.