Telegraf errors with non-nonsensical line number and message

Telegraf
1

I’m running telegraph 1.25.0 on rPis using RaspberryPi OS (bullseye) and without any changes to config files, except upgrade from 1.24 to 1.25 I’m suddenly getting the following error:

> /usr/bin/telegraf --config /etc/telegraf/telegraf.conf --config-direc
tory /etc/telegraf/telegraf.d --test
2023-01-05T19:16:31Z E! [telegraf] Error running agent: error loading config file /etc/telegraf/telegraf.conf: plugin inputs.socketstat: line 123: configuration specified the fields ["socket_types"], but they weren't used

So far so good, except that:

> tail +120 /etc/telegraf/telegraf.conf | head -n 10


# Read metrics about cpu usage
[[inputs.cpu]]
  ## Whether to report per-cpu stats or not
  percpu = false
  ## Whether to report total system cpu stats or not
  totalcpu = true
  ## If true, collect raw CPU time metrics
  collect_cpu_time = false

But I do use socketstat:

> egrep -n socketstat /etc/telegraf/telegraf.conf
546:[[inputs.socketstat]]

Line 123 vs. 546?
Also I don’t see what’s wrong with the config:

> tail +538 /etc/telegraf/telegraf.conf | head -n 20
[[inputs.procstat]]
  cgroup = "system.slice/caddy.service"
  pid_finder = "native"
  fieldpass = ["cpu_time_*", "memory_*", "*_usage"]
  tagexclude = ["cgroup_full", "user"]


# Gather indicators from established connections, using iproute2's ss command.
[[inputs.socketstat]]
  ## ss can display information about tcp, udp, raw, unix, packet, dccp and sctp sockets
  ## Specify here the types you want to gather
  socket_types = [ "tcp", "udp" ]
  ## The default timeout of 1s for ss execution can be overridden here:
  # timeout = "1s"


# Gather systemd units state
[[inputs.systemd_units]]
  ## Set timeout for systemctl execution
  # timeout = "1s"

I’m failing to see what’s wrong there, perhaps have been staring at it for too long and don’t see the obvious?

The full config file is available at https://tve.s3.amazonaws.com/public/telegraf.conf (I’m not allowed to upload here)

Appreciate any help! Thanks!

Update: the problem manifests itself with 1.24.3, 1.24.4, and 1.25.0. Everything works normally using 1.24.2. Packages pulled from:
Get:1 https://repos.influxdata.com/debian stable/main armhf telegraf armhf 1.24.2-1 [39.6 MB]

2

No idea why the weird line number, however the real reason is that the socketstat plugin does not recognize any “socket_types” option. The real name for the value is protocols seen here.

Looks like it has been this way since it was very first added feat: socketstat input plugin (#3649) · influxdata/telegraf@911c0a9 · GitHub as well!

3

Thanks for spotting the issue!! I hadn’t looked at the source code… Looks like the readme has been wrong since inception: feat: socketstat input plugin by sajoupa · Pull Request #3649 · influxdata/telegraf · GitHub

4

Would you be willing to file a bug? I would really appreciate it!

5