CloudWatch output doesn't send any metrics that has field type uint64

[wingyplus]

Relevant telegraf.conf:

[global_tags]

[agent]
  interval = "10s"
  round_interval = true
  metric_batch_size = 1000
  metric_buffer_limit = 10000
  collection_jitter = "0s"
  flush_interval = "10s"
  flush_jitter = "0s"
  precision = ""
  debug = true
  quiet = false
  logfile = "/var/log/telegraf/telegraf.log"
  hostname = "myhostname"
  omit_hostname = false

[[inputs.varnish]]
  name_suffix = "_cloudwatch"
  binary = "/usr/bin/varnishstat"
  stats = ["VBE.boot.frontend*.conn"]

[[outputs.cloudwatch]]
  region = "ap-southeast-1"
  namespace = "PWB/Varnish"
  namepass = ["*_cloudwatch"]

System info:

Telegraf version: 1.6.3-1
Operating system name: Debian
Kernel version: 4.9.0-6-amd64

Steps to reproduce:

1. starting telegraf with this configuration
2. testing telegraf config that can collect from input correctly

   $ sudo telegraf --config /etc/telegraf/telegraf.conf --test
   > varnish_cloudwatch,host=myhostname,section=VBE boot.frontend.pwb-magento-frontend-dev-6f7679603bb2bdb7.elb.ap-southeast-1.amazonaws.com(10.4.6.xx).conn=14i,boot.frontend.filtered.com(10.4.7.xx).conn=191i 1527833596000000000
   

3. wait until telegraf wrote metrics to CloudWatch

Expected behavior:

Should see metrics in namespace PWB/Varnish

Actual behavior:

Didn't see any metrics in namespace PWB/Varnish

Additional info:

After try to debug this problem. I found that varnish input try to parse number of field with uint64 (see https://github.com/influxdata/telegraf/blob/master/plugins/inputs/varnish/varnish.go#L144). But at cloudwatch output it doesn't handle type uint64 then become unsupport type (see https://github.com/influxdata/telegraf/blob/master/plugins/outputs/cloudwatch/cloudwatch.go#L173). That why it doesn't send any metrics to CloudWatch.

Currently, I fixed by handle uint64 type and cast it to float64 but I'm not sure it's the right way to do or not.