Odd behaviour with net.tcp.service

[bugjargal2]

Hi everyone,

On a brand new deployment of v 5.0 I encounter a very odd issue.

I have a bunch of hosts that I monitor using some common templates.
In one of the templates, I have the following item:

net.tcp.service[tcp,,5111]

All of them work as expected but one.

Here's a comparison between one host that works as expected and the faulty one. From the zabbix server :

The one that is ok:

$ zabbix_get -s 192.168.0.108 -k net.tcp.service[tcp,,5111]
1

$ telnet 192.168.0.108 5111
Trying 192.168.0.108...
Connected to 192.168.0.108.
Escape character is '^]'.

The one that has the issue:

$ zabbix_get -s 192.168.0.120 -k net.tcp.service[tcp,,5111]
0

telnet 192.168.0.120 5111
Trying 192.168.0.120...
Connected to 192.168.0.120.
Escape character is '^]'.

The two differences that I see between the faulty host and the others are:
-The one with issue runs the service on port 5111 in a docker container, the others natively in the host's OS
-The one that has the issue takes up to 2 seconds to display the telnet response while others are instant (but zabbix_get response is instantaneous for all of them).

Any pointer is very welcome.

Thanks