What the monitoring agent needs to be Healthy? ("ping problem")

[Danixu]

Hello,

Recently I have installed OpenWRT on my five devices (I was waiting for a management software like yours). On all the devices I have installed the Config Agent with the Luci addon to manage it via web, and I have configured the devices using the Organization Seret Key.
All the devices have joined the OpenWisp controller automatially without any problem, and I have installed the openwisp-agent on it to also have metrics and monitor all the devices.
Once the Agent is installed, the devices becomes Healthy and start sending metrics to the controler, but a few minutes later they becomes "CRITICAL". All the metrics still coming to the controller so the monitoring agent is working, but the Openwisp controller says that "ping is failing".
The devices are in the same network and there is no firewall to limit the traffic. I have also tested to ping between the devices and the controller, and is working in both ways. I have not seen any error in the logs folder so I am not sure what can be the problem.

The controller was installed following the Ansible instructions into a Debian 12 machine. My devices are using OpenWRT 23.05.4 (two), and latest nighty (three), and I have installed the app from the software manager (opkg). They are two different models so isn't related with the model.

Another thing I have noticed is that looks like the Monitoring entries doesn't have any data, even when the devices are sending the data to the controller and you can even see it in every device configuration menu without problem.

Memory in Monitoring:

Memory in Devices configuration:

The metrics are coming and stored into the device, but looks like they are not shown in the metrics setting.

Best regards.

[nemesifier]

Hi @Danixu,

Ping checks are performed by the server, the agent is not involved here.

Please read the following sections to understand how this mechanism works:

• Setting up the management network
• Passive vs active metric collection

Regarding the problem observed in the metric section, what URL is that?
It's possible there may be a bug, as we generally don't really use that section except for debugging purposes. But if you can share more detailed info we'll look into it and open an issue if there's any bug.

[Danixu]

Hello,

Thanks, I'll take a look to the links to see if I can determine why is failing, because there is connectivity from the controller to all the devices (are in the same network without firewall).
About the metrics, looks like it was too soon because now that some days has passed, the data is starting to appear.

Best regards.

[bmcgonag]

I'm also seeing this issue, but no idea where to set the value you linked for this.

OPENWISP_CONTROLLER_MANAGEMENT_IP_ONLY=false

Can you clarify where this is set? Is this an environment variable that needs to be set on the server?

[Danixu]

I have not checked it yet, but looks like they must be changed in the Django settings.py file:

https://openwisp.io/docs/dev/controller/user/settings.html#openwisp-controller-management-ip-only

Now I don't have access to the instance to test it, but I'll do as soon I can.

[nemesifier]

https://openwisp.io/docs/dev/user/django-settings.html

[bmcgonag]

Ok, put that into the ansible playbook under the variables section, and set to false. Do I need to do anything besides re-running the play?

[nemesifier]

@bmcgonag just run the playbook, if you're using the latest master of ansible-openwisp2, you can use the django_settings tag (-t django_settings) to run only the tasks which deal with this change.

[Danixu]

Finally the Django setting has not worked for me, but I have edited the configuration file manually just testing and it worked.

Looks like the luci-app-openwisp package used to configure the configuration agent is not getting the network name correctly. The name it gets is "lan" related with the network interface instead the device, whic it's real name is "br-lan". I don't know if it's a bug or something related with my setup.

config controller 'http'
        ...
        option management_interface 'br-lan'

@bmcgonag check if maybe is your problem too. Just connect to the router, run the ifconfig command to get the real interface name, and configure that line in the /etc/config/openwisp file.

Best regards.

[nemesifier]

After editing the django setting, did you restart the application?

[Danixu]

After editing the django setting, did you restart the application?

The entire server to be sure, but like the other way has worked I have not tried to investigate the problem.