We have setup watcher on two independent clusters, but these two are also managed via tribe nodes. I am wondering if watcher is supported or installed on the tribe node(s) to begin with. If so, does all of the configuration for watcher then reside in the elasticsearch.yml for the tribe nodes at the "root" (i.e. verbatim what I currently have in elasticsearch.yml on the cluster nodes) or does it need to be nested under each cluster managed by the tribe (for example: tribe.mastering_one.watcher.actions.email.)?
In your situation, I suggest installing Watcher on just one of the two clusters - much less complicated, easier to query watch history, etc.
One of the nice things about Watcher is that it can easily make requests to external clusters (or really, anything with a RESTful API, including your tribe Node(s)). We describe how to do this in the Watcher Docs .
Thanks Mark & Steve. It is easier to install on 1 cluster and query over to the other via the http input. Although, we may opt to run watcher in each cluster separately. Appreciate the feedback.
Apache, Apache Lucene, Apache Hadoop, Hadoop, HDFS and the yellow elephant
logo are trademarks of the
Apache Software Foundation
in the United States and/or other countries.