We are monitoring our Jenkins environment with a Prometheus server. On the Jenkins side we are using the Prometheus Metrics plugin (version 2.0.10). On the Prometheus side there is a strange error message on one of our endpoint:

"label name "node" is not unique: invalid sample"

And the endpoint is marked as "DOWN". After checking the $JENKINS_URL/prometheus page, the problem is obvious:

1. TYPE jenkins_node_buildbot1_jenkins_node_builds summary
   jenkins_node_builds{node="buildbot1_jenkins_node",node="master",quantile="0.5",} 1139.0694669270001
   ...
2. TYPE jenkins_node_buildbot2_jenkins_node_builds summary
   jenkins_node_builds{node="buildbot2_jenkins_node",node="master",quantile="0.5",} 1.569607388
   ...
3. TYPE jenkins_node_buildbot3_jenkins_node_builds summary
   jenkins_node_builds{node="buildbot3_jenkins_node",node="master",quantile="0.5",} 1.657462223
   ...

The node label is duplicated.