Please provide verification steps for this one.

Steps to reproduce:

1. Set up metrics on the engine and hosts
2. Check the records in Kibana - 
     The plugin converts collectd fields to be nested under "collectd".
     The numeric value of the record is nested under "collectd"-> "plugin" -> "type". 
     A record can have one or more numeric values.

Previously this was done not in a dedicated plugin but in the config file.

Examples of expected results:

"collectd": {
      "interval": 10,
      "plugin": "processes",
      "type": "ps_state",
      "type_instance": "sleeping",
      "processes": {
        "ps_state": 272
      }
    }, 

"collectd": {
      "interval": 10,
      "plugin": "disk",
      "plugin_instance": "sda",
      "type": "disk_octets",
      "disk": {
        "disk_octets": {
          "read": 0,
          "write": 1899692.00845666
        }
      }

(In reply to Shirly Radco from comment #2)
> Steps to reproduce:
> 
> 1. Set up metrics on the engine and hosts
> 2. Check the records in Kibana - 
>      The plugin converts collectd fields to be nested under "collectd".
>      The numeric value of the record is nested under "collectd"-> "plugin"
> -> "type". 
You sure about this?
>      A record can have one or more numeric values.
> 
> Previously this was done not in a dedicated plugin but in the config file.
> 
> Examples of expected results:
> 
> "collectd": {
>       "interval": 10,
>       "plugin": "processes",
>       "type": "ps_state",
>       "type_instance": "sleeping",
>       "processes": {
>         "ps_state": 272
IIUC this is under collectd->processes->ps_state and not collectd->plugin->type
>       }
>     }, 
> 
> "collectd": {
>       "interval": 10,
>       "plugin": "disk",
>       "plugin_instance": "sda",
>       "type": "disk_octets",
>       "disk": {
>         "disk_octets": {
>           "read": 0,
>           "write": 1899692.00845666

IIUC this is under collectd->disk->disk_octets->write and not collectd->plugin->type
>         }
>       }

Thus what am I really looking for, the data are indeed nested under collectd however does not fit your specification

example output on my machine:

{
  "_index": "ovirt-metrics-2017.06.20",
  "_type": "fluentd",
  "_id": "AVzEtQeqCiT5_6MSsdhE",
  "_score": 1,
  "_source": {
    "hostname": "******************************",
    "collectd": {
      "dstypes": [
        "gauge"
      ],
      "interval": 10,
      "plugin": "memory",
      "type": "memory",
      "type_instance": "buffered",
      "memory": {
        "memory": 2195456
      }
    },
    "tag": "project.ovirt-metrics-engine",
    "ovirt": {
      "entity": "engine"
    },
    "ipaddr4": "***********",
    "ipaddr6": "*****************************",
    "@timestamp": "2017-06-20T10:35:49+02:00"
  }
}

Is this correct they are nested under collectd?
What did you mean by 'The numeric value of the record is nested under "collectd"-> "plugin" -> "type".'?

This is ok.
This means the field name is built as 

"collectd" 
    { "memory"(plugin_field) 
        { "memory" (type_field) :2195456
        }
    }

I see, thank you, due to comment #3 and #4 moving to verified