Hi, Quick question w.r.t the health API that is available in the cluster operations, is it supposed to give information at a node level or at a cluster level. The documentation mentions that it is for a typesense node, but when we have a 3 node self hosted cluster with 2 nodes down and if we try to hit the API on the node that has the process running we are getting the response that the health is not ok and once we bring one more node up (i.e 2 nodes up and 1 node down), the same health response is coming as ok.

The health end-point refers to node health. But 3-node cluster can survive only 1 node failure. If 2 of the nodes are down, even the third node will return unhealthy because Raft requires a quorum of nodes available.

So in such a case is there some other path without the authorisation headers that we could use to check if the process is up and running on a node. This might be needed when we need to setup the health probes for k8's, etc.

Just use a tcp port check instead of http check.

Sure this was my last resort, was hoping if there was any other way. Thanks