Elasticsearch Metrics are not correctly parsed/emitted to OpenTSDB

[langerma]

Relevant telegraf.conf:

[agent]
  interval = "10s"
  round_interval = true
  metric_batch_size = 100
  metric_buffer_limit = 4000000
  collection_jitter = "0s"
  flush_interval = "10s"
  flush_jitter = "0s"
  precision = ""
  debug = false
  quiet = false
  hostname = ""
  omit_hostname = false

[[outputs.opentsdb]]
   host = "http://127.0.0.1"
   port = 14242
   httpBatchSize = 100

[[inputs.elasticsearch]]
   servers = ["http://localhost:9200"]
   http_timeout = "5s"
   local = true
   cluster_health = true
   cluster_stats = true

System info:

telegraf-1.7.4-1.x86_64.rpm / RHEL 7 / Elasticsearch 6.4

Steps to reproduce:

1. install telegraf 1.7.4-1
2. restart telegraf

Expected behavior:

same config works with telegraf 1.5.3-1.x86_64.rpm everything works fine

Actual behavior:

10:16:10.310 WARN [PutDataPointRpc.execute] - Invalid tag value: empty string: metric=elasticsearch_clusterstats_indices_segments_count ts=1536653770 value=12156 cluster_name=george-visualisation environment=PROD host=gevismasterp1at.eb.lan.at node_name= status=green

this error is the same for all clusterstats metrics.

[glinton]

Thanks for the report.

[langerma]

any updates on this issue?

[glinton]

@langerma did you update your version of opentsdb or another part of your stack while updating to telegraf 1.7.4? Nothing really changed in the tsdb output since 1.5.3 and I wasn't able to find anywhere in the elastic input plugin that would cause the node_name to be empty. I'm continuing to investigate however.

[langerma]

no i did no update...neither on opentsdb side nor elastic side. i guess it is a problem with the elastic input.

[langerma]

thx, will try once the new telegraf version is released.