Error in Netty pipeline: java.io.IOException: Connection reset by peer

[smm (Stefano)]

Dear community,
in my logstash 7.8.1 logs I get this error message quite a lot:
[2021-09-07T08:36:56,031][ERROR][logstash.inputs.tcp ][main] Error in Netty pipeline: java.io.IOException: Connection reset by peer
It only occurs when data is sent from google/kubernetes.
Also: In these cases we have data loss.

Do you know a reason and / or a workaround for this?

Our input plugin part looks like this:

input {

  redis {
    host => "127.0.0.1"
    data_type => "list"
    key => "logstash"
    codec => json
    threads => 8
    add_field => { "[@metadata][datatype]" => "log" }
  }

  redis {
    host => "127.0.0.1"
    data_type => "list"
    key => "filebeat"
    codec => json
    threads => 6
    add_field => { "[@metadata][datatype]" => "log" }
  }

  redis {
    host => "127.0.0.1"
    data_type => "list"
    key => "winlogbeat"
    codec => json
    threads => 1
    add_field => { "[@metadata][datatype]" => "log" }
  }

  redis {
    host => "127.0.0.1"
    data_type => "list"
    key => "beat"
    codec => json
    threads => 4
    add_field => { "[@metadata][datatype]" => "beat" }
  }

  beats {
    id => "beats-input"
    port => 5044
    add_field => { "[@metadata][indexprefix]" => "%{[@metadata][beat]}-%{[@metadata][version]}" }
  }
  
  tcp {
    port => 5518
    codec => "fluent"
    tcp_keep_alive => true
    tags => ['basic','docker','json']
  }
}

Could be some kind of timeout / too much data problem.

Thanks for any insight you can give in this!
Cheers