Problem configuration telegraf

[Matthieubmt]

Hello,

I encountered an issue while attempting to incorporate your code into my Telegraf configuration.

'error loading config file /etc/telegraf/telegraf.conf: plugin inputs.exec: line 3793: configuration specified the fields ["timestamp_key"], but they weren't used'

Can you please advise on the possible cause of this problem?

[XReyRobert]

Hi there, no idea, you can try

/usr/bin/env bash -c '/usr/bin/nvidia-smi -q -x | /usr/local/bin/xq .

make sure that you have all the required command and that the "nvidia_smi_log.timestamp" key is present in the output...

[Yomisana]

Hi there 👋

I think telegraf example conf.

[[inputs.exec.json_v2]]
  measurement_name = "nvidia-smi"
  #timestamp_key= "nvidia_smi_log.timestamp" <<< I looked at similar docs it put in json_v2.object but it still doesn't put data into influxDB for me

I check this: https://github.com/influxdata/telegraf/blob/release-1.13/plugins/inputs/nvidia_smi/README.md

[XReyRobert]

this is my currently running configuration...

[[inputs.exec]]
  ## Commands array
  commands = ["/usr/bin/env bash -c '/usr/bin/nvidia-smi -q -x | /usr/local/bin/xq .'"]
  data_format = "json_v2"

[[inputs.exec.json_v2]]
  measurement_name = "nvidia-smi"

[[inputs.exec.json_v2.tag]]
  path = "nvidia_smi_log.gpu.@id"
  rename = "gpuid"

[[inputs.exec.json_v2.object]]
  timestamp_key= "nvidia_smi_log.timestamp"
  path = "nvidia_smi_log"
  excluded_keys = ["gpu_processes"]

[Yomisana]

Stable Diffusion Public Plane Just in case someone says the control panel is unusable... In Control Panel>Settings>variable

• There are two items that need to be set to the corresponding bucket and the corresponding Data source.
• Then go back to the visual control panel and edit every put the similar

  For me is:
  from(bucket: "nvidia-smi")

  Change it to your own storage bucket name and check out right side "Fields" choose your current gpu. (In normal... You don't need edit "Fields" but sometime you should edit it...) :shipit:

[gcs8]

For people running into telegraf barfing at this, roll yq back to version 3.2.1 and it should work again.

pip uninstall yq pip install yq==3.2.1

I left a complaint here, if someone who is better at code wants to go add something of value over there. https://github.com/kislyuk/yq/issues/177