Firstly I should mention that I already know about the whole xpack.monitoring.cluster_uuid thing that was fixed in this PR. However this doesn't seem to be the case here.
I have a production cluster, which is being monitored by a separate monitoring cluster as suggested in the docs. There are 3 instances of logstash on three servers, which are configured exactly the same, have the same pipelines exposed on the same ports... everything is the same! And indeed all of the logstash pipelines have the same elasticsearch output, so they should report the cluster_uuid the same.
The abnormality is that "only one pipeline" from "only one server" doesn't report the cluster_uuid, and is being categorized in the famous "Standalone Cluster". Hence in the monitoring cluster I have 2 clusters! The correct one has all the pipelines, showing 3 nodes for each pipeline "except for the pipeline I mentioned", and one other "Standalone cluster" with only one pipeline with a single node in it.
I guess there must be a bug somewhere, since as I mentioned, every single pipeline config file has the same elasticsearch output. How can this be fixed? or at least how can I pin point the problem?