Describe the bug Impossible to start the telegraf serivce Hav

Well, I have: all.yml <div clas

Hi <a class="user-mention notranslate" data-hovercard-type="user" data-hovercard-url="

Daemon crash because of empty config.conf and plugin.conf files. about ansible-telegraf HOT 3 CLOSED

dj-wasabi commented on May 31, 2024

Daemon crash because of empty config.conf and plugin.conf files.

from ansible-telegraf.

Comments (3)

dj-wasabi commented on May 31, 2024

Hi,

I tried to replicate this.
I have the following configuration set into my Molecule config:

  inventory:
    group_vars:
      all:
        telegraf_plugins_extra_exclusive: True
        telegraf_plugins_extra:
          nginx:
            config:
              - urls = ["http://localhost:9090/nginx_status"]
          phpfpm:
            config:
              - urls = ["http://localhost:9090/fpm-status"]
        telegraf_agent_output:
          - type: influxdb
            config:
              - urls = ["http://influxdb:8086"]
              - database = "telegraf"
              - precision = "s"
        telegraf_plugins_extra:
          percpu-usage:
            plugin: cpu
            config:
              - percpu = true
              - totalcpu = false
              - name_override = "percpu_usage"
              - fielddrop = ["cpu_time*"]

No other config, only the above and what is provided in the default/main.yml. But I don't see any empty files. Do you have some other configuration set somewhere?

Kind regards,
Werner

from ansible-telegraf.

tristanbes commented on May 31, 2024

Well, I have:

all.yml

telegraf_agent_output:
  - type: influxdb
    config:
      - urls = ["http://influx.foo.com:8086"]
      - database = "telegraf"
      - username = "{{ influx_user }}"
      - password = "{{ influx_password }}"

telegraf_plugins_extra: {}
telegraf_plugins_extra_exclusive: True

web.yml

## Monitoring
telegraf_plugins_extra:
  nginx:
    config:
      - urls = ["http://localhost:9090/nginx_status"]
  phpfpm:
    config:
      - urls = ["http://localhost:9090/fpm-status"]
  varnish: ~

But I've been using your playbook since ages ago, so maybe it was due to consecutive updates ?

from ansible-telegraf.

dj-wasabi commented on May 31, 2024

Hi @tristanbes

I was able to replicate it. I created a PR for when the property telegraf_plugins_extra and telegraf_plugins_default are empty (then Telegraf won't start) I enable default the "mem" plugin so it won't crash/stop working. Otherwise I have no idea to prevent it.

If you or some one else has a better suggestion, then please let me know and I'll implement it.

from ansible-telegraf.

Recommend Projects