Fix isMaster checking while getting node stats

[vbeskrovnov]

Bug report

Relevant telegraf.conf:

[[inputs.elasticsearch]]
   cluster_health = true

System info:

Telegraf v.1.2.1, RedHat

Steps to reproduce:

1. Deploy elasticsearch cluster with 2 or more nodes
2. Configure telegraf to work with cluster
3. Set cluster_health in telegraf.conf to true
4. Run telegraf

Expected behavior:

Information about count of master and data nodes always writing in elasticsearch_clusterstats_nodes.

Actual behavior:

Information about count of master and data nodes not always writing in elasticsearch_clusterstats_nodes.

Additional info:

This part of code executed in cycle for every node in cluster, so information only about last node saved in e.isMaster. It can be reproduced via tests by adding second node with id different from master node in nodeStatsResponse in test data.

if e.ClusterStats {
 	// check for master
 	e.isMaster = (id == e.catMasterResponseTokens[0])
 }

[danielnelson]

This definitely looks like a bug to me, but I don't think your fix will solve it. The queries to the servers are done concurrently and it is anyones guess what the value will be. We shouldn't be using a field on the shared struct here at all.