influxdb_v2 output not respecting rate limiting response

[russorat]

Relevant telegraf.conf:

[[outputs.influxdb_v2]]
  ## The URLs of the InfluxDB cluster nodes.
  ##
  ## Multiple URLs can be specified for a single cluster, only ONE of the
  ## urls will be written to each interval.
  ## urls exp: http://127.0.0.1:9999
  urls = ["https://us-west-2-1.aws.cloud2.influxdata.com"]

  ## Token for authentication.
  token = "TOKEN"

  ## Organization is the name of the organization you wish to write to; must exist.
  organization = "ORGID"

  ## Destination bucket to write into.
  bucket = "BUCKETID" 

System info:

local telegraf, influxdb cloud2 instance.

Steps to reproduce:

1. sign up for a free influxdb cloud 2 account
2. attempt to write many metrics into the account in order to get a rate limit on cardinality
3. notice that the message does not have the correct retry information

Expected behavior:

the message in the telegraf logs should have the correct retry time.

Actual behavior:

the message in the telegraf logs always has 0s.

Additional info:

2019-09-10T16:05:04Z E! [outputs.influxdb_v2] when writing to [https://us-west-2-1.aws.cloud2.influxdata.com]: Waiting 0s for server before sending metric again

[glinton]

Can you make a write request using curl and use the -i flag? Telegraf uses a header in the response, I wonder if it changed without us being notified