NetXMS Support Forum

This works from the NetXMS server console:


>nxwsget 127.0.0.1 http://10.9.8.7/metrics /started
/started = 1760481688751


Neither /started nor $.started work (with 'metrics' as the WS Name and using the URL above):

>nxget -D 6 127.0.0.1 'metrics:/started'
[crypto             ] Validating ciphers
[crypto             ]    AES-256 enabled
[crypto             ]    Blowfish-256 disabled (validation failed)
[crypto             ]    IDEA disabled (config)
[crypto             ]    3DES enabled
[crypto             ]    AES-128 enabled
[crypto             ]    Blowfish-128 disabled (validation failed)
[crypto             ] Crypto library initialized (OpenSSL 3.0.13 30 Jan 2024)
[agent.conn.1       ] New connection created (address=127.0.0.1 port=4700 compression=allowed)
[agent.conn.1       ] Using NXCP version 5
[agent.conn.1       ] Sending message CMD_REQUEST_SESSION_KEY (1) to agent at 127.0.0.1
[agent.conn.1       ] Received message CMD_SESSION_KEY (1) from agent at 127.0.0.1
[agent.conn.1       ] Sending message CMD_SET_SERVER_CAPABILITIES (2) to agent at 127.0.0.1
[agent.conn.1       ] Received message CMD_REQUEST_COMPLETED (2) from agent at 127.0.0.1
[agent.conn.1       ] Sending message CMD_GET_PARAMETER (3) to agent at 127.0.0.1
[agent.conn.1       ] Received message CMD_REQUEST_COMPLETED (3) from agent at 127.0.0.1
[agent.conn.1       ] Sending message CMD_GET_TABLE (4) to agent at 127.0.0.1
[agent.conn.1       ] Received message CMD_REQUEST_COMPLETED (4) from agent at 127.0.0.1
[agent.conn.1       ] Sending message CMD_GET_LIST (5) to agent at 127.0.0.1
[agent.conn.1       ] Received message CMD_REQUEST_COMPLETED (5) from agent at 127.0.0.1
404: Unknown metric
[agent.conn.1       ] Receiver loop terminated
[agent.conn.1       ] Receiver cleanup completed


What am I doing wrong?

I have configured a remote agent trough the agent tunnel and linked the agent to a container in a zone. I have setup network discovery for the zone and selected the agent as the proxy node. In the agent logs, I can see the scan is being activated and lots of devices are found with both icmp and snmp scan, but nodes are not being added. Anyone who can guide me where to look for issues?

Здравствуйте,
Использую в работе агенты с включенным режимом прокси для мониторинга устройств в удаленных офисах. При этом если теряется контакт с прокси нодой на более чем 5 мин (время настроено через таймер) то уведомление приходит только от прокси ноды (тут все как и ожидалось). Когда контакт с прокси нодой восстанавливается то уведомления приходят и от прокси ноды и от подчиненных нод (тоже как и ожидалось).
Проблема появляется когда контакт с прокси нодой теряется кратковременно (т.е. контакт потерян и восстановлен в пределах 5 мин). В таком случае уведомление от прокси ноды не приходит (как и ожидалось) но зато приходят уведомления от всех подчиненных нод что они появились в сети. И так каждый раз при кратковременной потери контакта с прокси нодой.
Как этого можно избежать?

Hello,

How many alarms do you have in NetXMS? You can do "sh st" command in server debug console to see the number. In 5.2 it worked that way that if Alarms tab was opened at some moment, it stayed in background and consumed resources. This will be fixed in 5.3

One recent finding is that Stacked line charts can lead to behavior like this - this is currently being fixed.

And you can check .nxmc4/nxmc.log file in your home folder for any errors

.apkg installer is actually a shell script joined with archive that contains agent file. If you do
head nxagent-5.2.6-linux-x86_64.apkg -n 118
you'll see the contents of the script.

I've tried to install it Ubuntu 24 Server, did not get the result that you are observing.

.apkg installer writes log file to /tmp/nxagentupdate.log. If you can repeat the installation, can you share that file?

Well, yes, FindAlarmByKey() is not searching through already terminated alarms, only current ones. And even if everything happens within one  EPP rule, alarm termination happens first and server action happens after that.

So the solution for now is to store the ID somehow before alarm termination. We can use filtering script in the EPP rule:

a = FindAlarmByKey($event.lastAlarmKey);
$event.setNamedParameter("lastAlarmID", a.id);
return true;
return true; is needed since this is filtering script and EPP rule will be processed further only if this script returns true.

And in your script that is used by the %[script] macro:
return $event.getParameter("lastAlarmID");

We are adding a new parameter to the event. This parameter will be accessible in any EPP rule that goes below the one where it's being added.

Ah, and actually you don't need the %[script] thing any more - you can just use %<lastAlarmID> macro - this should return value of event's parameter.

Hi Filipp,

thanks for the reply that also works for new Alarms but this issue I have is when the event is raised for terminating the alarm it doesn't work and raises the below

Error:
Minor   vm-netxms-srev   SYS_SCRIPT_ERROR   Script (getAlarmID) execution error: Error 14 in line 3: Function or operation argument is not an object   07.10.2025 08:58:16

Script:
//return $event.id;
a = FindAlarmByKey($event.lastAlarmKey);
return a.id;

I concur, this is the way to go. I'm going to ask the NetXMS team for their input. The company I work for has a support contract. Upgrades would be so simple if we could get this.

Currently it only writable by jira or redmine drivers, so only by developing a new driver.
But in overall it would be good to allow setting it from nxsl (because integration now can be done just with nxsl and web services). I've created ticket for that: https://track.radensolutions.com/issue/NX-2843