MQTT consumer plugin crashes telegraf on a connection failure

[moracca]

Relevant telegraf.conf

[agent]
  interval = "10s"
  round_interval = true
  metric_batch_size = 1000
  metric_buffer_limit = 10000
  collection_jitter = "0s"
  flush_interval = "10s"
  flush_jitter = "0s"
  precision = "0s"
  hostname = ""
  omit_hostname = false

[[inputs.mqtt_consumer]]
  servers = ["tcp://172.30.88.239:1883"]

[[inputs.ping]]
  urls = ["8.8.8.8"]
  count = 6
  ping_interval = 0.1
  timeout = 2.00

[[outputs.file]]
  files = ["stdout"]

Logs from Telegraf

My-Mac-Pro:telegraf ldd$ telegraf --config full.conf --config-directory telegraf.d/
2023-04-04T06:42:32Z I! Loading config file: full.conf
2023-04-04T06:42:32Z I! Loading config file: telegraf.d/outputs.file.conf
2023-04-04T06:42:32Z I! Loading config file: telegraf.d/outputs.postgresql.conf
2023-04-04T06:42:32Z I! Starting Telegraf 1.26.0
2023-04-04T06:42:32Z I! Available plugins: 235 inputs, 9 aggregators, 27 processors, 22 parsers, 57 outputs, 2 secret-stores
2023-04-04T06:42:32Z I! Loaded inputs: cisco_telemetry_mdt disk mqtt_consumer ping (2x)
2023-04-04T06:42:32Z I! Loaded aggregators: 
2023-04-04T06:42:32Z I! Loaded processors: regex
2023-04-04T06:42:32Z I! Loaded secretstores: 
2023-04-04T06:42:32Z I! Loaded outputs: file postgresql
2023-04-04T06:42:32Z I! Tags enabled: host=My-Mac-Pro.local
2023-04-04T06:42:32Z I! [agent] Config: Interval:10s, Quiet:false, Hostname:"My-Mac-Pro.local", Flush Interval:10s
2023-04-04T06:43:02Z E! [telegraf] Error running agent: starting input inputs.mqtt_consumer: network Error : dial tcp 172.30.88.239:1883: i/o timeout
My-Mac-Pro:telegraf ldd$ 

### System info

Telegraf 1.26.0 on FreeBSD 13.1

### Docker

_No response_

### Steps to reproduce

run telegraf with the attached config, and it will exit after timing out

### Expected behavior

as per a discussion in slack:

Plug-ins that connect to external services should retry instead of crashing telegraf. If you believe this is not the case with the mqtt plugin, please file a bug report.

The telegraf process should continue to function for the other inputs (ping), and the mqtt plugin should retry its connection periodically

### Actual behavior

process exits after the i/o timeout

### Additional info

_No response_

[Hipska]

Hi, it seems like this is already discussed in #10694

[powersj]

Agreed - closing as a dup of #10694 - I'll rename that one to call out MQTT