InfluxDB Writes 0 metrics

[s1m0]

I have several [[inputs.exec]] sections that telegraf runs every 5s to 5m. When an exec isn't run, telegraf still writes an empty http POST to InfluxDB. See example logging output below; there are 8 [[outputs.influxdb]] sections writing to 4 different databases on 2 separate installations (a single test server, migrating to a 3-way production cluster). Every 5 seconds 28 or 29 metrics are written to the 2 InfluxDB test and production databases but the others don't run so 0 metrics are available to be written, yet the empty message is still sent.

2016/03/17 16:25:25 Gathered metrics, (separate 5s interval), from exec in 74.880001ms
2016/03/17 16:25:25 Wrote 0 metrics to output influxdb in 474.396µs
2016/03/17 16:25:25 Wrote 0 metrics to output influxdb in 634.736µs
2016/03/17 16:25:25 Wrote 0 metrics to output influxdb in 898.01µs
2016/03/17 16:25:25 Wrote 0 metrics to output influxdb in 1.008772ms
2016/03/17 16:25:25 Wrote 0 metrics to output influxdb in 1.491286ms
2016/03/17 16:25:25 Wrote 0 metrics to output influxdb in 1.332947ms
2016/03/17 16:25:25 Wrote 28 metrics to output influxdb in 3.004514ms
2016/03/17 16:25:25 Wrote 28 metrics to output influxdb in 2.784524ms
2016/03/17 16:25:30 Gathered metrics, (separate 30s interval), from exec in 156.797282ms
2016/03/17 16:25:30 Gathered metrics, (separate 5s interval), from exec in 176.83572ms
2016/03/17 16:25:30 Wrote 29 metrics to output influxdb in 5.687058ms
2016/03/17 16:25:30 Wrote 0 metrics to output influxdb in 2.706675ms
2016/03/17 16:25:30 Wrote 28 metrics to output influxdb in 6.591603ms
2016/03/17 16:25:30 Wrote 0 metrics to output influxdb in 3.17213ms
2016/03/17 16:25:30 Gathered metrics, (separate 30s interval), from exec in 213.787362ms
2016/03/17 16:25:30 Wrote 0 metrics to output influxdb in 6.457787ms
2016/03/17 16:25:30 Wrote 0 metrics to output influxdb in 7.900459ms
2016/03/17 16:25:30 Wrote 29 metrics to output influxdb in 8.945402ms
2016/03/17 16:25:30 Wrote 28 metrics to output influxdb in 5.779109ms
2016/03/17 16:25:30 Gathered metrics, (separate 5m0s interval), from exec in 522.519729ms
2016/03/17 16:25:31 Gathered metrics, (separate 1m0s interval), from exec in 1.300327534s
2016/03/17 16:25:35 Gathered metrics, (separate 5s interval), from exec in 78.649291ms
2016/03/17 16:25:35 Wrote 1 metrics to output influxdb in 4.443708ms
2016/03/17 16:25:35 Wrote 5 metrics to output influxdb in 3.774689ms
2016/03/17 16:25:35 Wrote 28 metrics to output influxdb in 5.265983ms
2016/03/17 16:25:35 Wrote 12 metrics to output influxdb in 4.738119ms
2016/03/17 16:25:35 Wrote 5 metrics to output influxdb in 137.146364ms
2016/03/17 16:25:35 Wrote 12 metrics to output influxdb in 137.678064ms
2016/03/17 16:25:35 Wrote 1 metrics to output influxdb in 138.070356ms
2016/03/17 16:25:35 Wrote 28 metrics to output influxdb in 138.275571ms

This is a sample of the InfluxDB logging showing the empty messages being logged (I tagged each influxdb output section with a different user_agent string so I could track the messages. In this case OSS-FC & OSS-O:

[httpd] 2016/03/18 00:01:15 write body received by handler:
[httpd] 2016/03/18 00:01:15 write body received by handler:
[http] 2016/03/18 00:01:15 127.0.0.1 - - [18/Mar/2016:00:01:15 +0000] POST /write?consistency=&db=FlexCache&precision=s&rp=default HTTP/1.1 204 0 - OSS-FC 88649365-ec9c-11e5-aee6-000000000000 130.237µs
[http] 2016/03/18 00:01:15 127.0.0.1 - - [18/Mar/2016:00:01:15 +0000] POST /write?consistency=&db=Origin&precision=s&rp=default HTTP/1.1 204 0 - OSS-O 88649707-ec9c-11e5-aee7-000000000000 72.56µs

By the way, what is the "consistency=" option?

[sparrc]

agreed, we should skip when there are 0