Monitoring TCP Connection aborts

Hi all,
I have to monitor, if a tcp connection between 2 servers in 2 different network segments (separated by firewalls) is stabil or gets aborts after a couple of time. So the 2 server should have a permanent connection, which is monitored. Is it correct, that a Icinga master has a permanent connection to his satellites, which is checked continuously by the cluster (or cluster-zone) check? My first idea was, just to install a Icinga Satellite in the other network segment and use the cluster checks to see aborts.

Anyway, are you guys are monitoring sth. like this, too and has a best practice?


I would say the cluster check for a satellite zone is a good option to monitor this.
I use the cluster check command to verify the connection of my satellites to the master and run them every 30s.
But I also do some simple check_tcp or icmp checks to endpoints/IPs in the remote network.

1 Like

Just found this by @Napsty :
Which gives Perfdata of e.g.

Maybe this is better?

The plugins by Claudio Kuenzler are usually very good, but I don’t think this check does what you need:

check_netio (

Last update: February 13, 2020

This is a plugin to monitor network interfaces on Linux systems. The initial version was developed by Ian Yates. The plugin is currently maintained by Claudio Kuenzler.

The plugin needs to be executed on the same host as you want the interfaces monitored. A typical use case of this plugin is in combination with a remote plugin execution service (e.g. NRPE). The plugin shows performance data related to the selected interface (Bytes in/out, Errors in/out, Drops in/out) and optional system wide tcp related statistics, retrieved from the OS.

As of the current state, the plugin does not support thresholds.

That’s not a real problem. Using InfluxDB + Grafana gives me the possibility to create an Alert Rule.
And maybe Claudio is interested in expand it. Maybe, I can also expand it and PR it (if I have the time :smiley: )