Make a chain/proxy by telegraf

Telegraf
, ,
1

Hello! I have:

  • Host A with InfluxDB
  • Host B with Telegraf and straight connection to Host A
  • Host C with Telegraf and straight connection to Host B (but not A)
    How can I set up Host B telegraf to send metrics from Host C to InfluxDB?

What I’ve done already:
[[inputs.influxdb_v2_listener]] on Host B and [[outputs.influxdb_v2]] on Host C
Host C doesn’t see any problem and sends metrics like normal. But I can’t see metrics with host=Host C in the Data Explorer. So I guess problem is in the Host B.
I tried to turn on debug on both hosts, but it didn’t help, no information about transmission.
There are no complete examples about inputs.influxdb_v2_listener

configs:
Host B

[[inputs.influxdb_v2_listener]]
service_address = "192.168.0.45:9086"

Host C

[agent]
interval = "5s"
round_interval = false
metric_batch_size = 8000
metric_buffer_limit = 1000000
collection_jitter = "0s"
flush_interval = "10s"
flush_jitter = "0s"
hostname = "$INFLUX_HOSTNAME" # Host C
omit_hostname = false

[[outputs.influxdb_v2]]
alias = "master_db"
urls = ["192.168.0.45:9086"]
token = "$INFLUX_TOKEN"
organization = "ORG"
exclude_bucket_tag = true
bucket = "master"
bucket_tag = "influxdb_database"
timeout = "120s"
content_encoding = "gzip"

Maybe someone have guesses or examples what I’m missing?

2

That looks completely correct…

Are you sure there is no metric filtering done on Host B?

3

yeah, i’m sure, all other inputs are the same on both Host B and Host C

4

I mean filtering on the processors/aggregators or outputs…

5

No, nothing. I don’t even use processors or aggregators.
Output on Host B is practically same:

[[outputs.influxdb_v2]]
    alias = "master_db"
    urls = ["$INFLUX_HOST"]
    token = "$INFLUX_TOKEN"
    organization = "ORG"
    exclude_bucket_tag = true
    bucket = "master"
    bucket_tag = "influxdb_database"
    timeout = "120s"
    content_encoding = "gzip"
6

And what is the content of the tag influxdb_database on metrics that come from Host C? Is it the same as for metrics on Host B?

7

yes, the same
It’s like this

[[inputs.swap]]
[[inputs.system]]
fielddrop = ["uptime_format"]
[[inputs.docker]]
endpoint = "unix:///var/run/docker.sock"
gather_services = false
source_tag = false
container_name_include = []
container_name_exclude = []
timeout = "10s"
total = false
docker_label_include = []
docker_label_exclude = []
tag_env = ["JAVA_HOME", "HEAP_SIZE"]
[inputs.docker.tags]
influxdb_database = "docker" # !!!!!!!!
8

If you want to keep the bucket for metrics coming from Host C, you will need to add the following to your inputs.influxdb_v2_listener on Host B:

bucket_tag = "influxdb_database"
9

Thx for your help, it was included, I pasted this part above.
The problem was with time on Host C, it was not syncing with ntp suddenly.

10

No, on the listener part on Host B, so like this:

[[inputs.influxdb_v2_listener]]
  service_address = "192.168.0.45:9086"
  bucket_tag = "influxdb_database"

But glad you figured it out…

11

You’re fantastic. Thank you again)

12

We needed to add the http:// to the urls in the [[outputs.influxdb_v2]] configuration block

urls = [“http://x.x.x.x:port”]
instead of the supplied example
urls = [“192.168.0.45:9086”]

But the proxy functionality worked fine