Generating Custom Juniper Syslog Messages

I wanted to focus on a lesser known feature which I’ve found useful over the years when trying to setup NMS alerting and logging which doesn’t typically garner much attention in the documentation.

This would be the logger utility available on any JunOS platform from the shell. This program allows you to generate virtually any syslog event at will to test your configurations and confirm you have the correct hosts, priority, facility and filtering set on your local host as well as confirm your alerting is working correctly w/o having to flap interfaces, or other intrusive testing.

Here’s a quick overview on our available Severity and Facility definitions.

So let’s say I want to generate a syslog message to a remote syslog collector, but I only want to send messages which are generated by the external facility with a severity of info or better.

This process is very well outlined in the Day One Book : Applying Junos Event Automation

But to save the hunting, here’s the relevant excerpt;


How to use the logger test utility

1. The logger utility is a shell command, and so the user must first start a system shell by invoking the start shell command:

    user@Junos> start shell %

2. The logger utility has the following command syntax: logger -e EVENT_ID -p SYSLOG_PRIORITY -d DAEMON -a ATTRIBUTE=VALUE MESSAGE. Only the EVENT_ID is required, and it must be entered entirely in uppercase:

    % logger -e UI_COMMIT

The above command causes a UI_COMMIT event to be generated, originated from the logger daemon, with no attributes, no message, and a syslog facility/severity of user/notice.

The default settings can be altered by using one of the optional command line arguments.

3. For an alternate syslog facility/severity use the -p argument and specify the facility/severity in the same facility.severity format used by the jcs:syslog() function:

    % logger -e UI_COMMIT –p

MORE? See Day One: Applying Junos Operations Automation for a table that lists the valid syslog facilities and severities for the jcs:syslog() function. 4. To alter what daemon generated the event, use the -d argument:

    % logger -e UI_COMMIT –d mgd

5. Include attributes for the event by using the -a argument. Use the argument multiple times if more than one attribute is needed. The attribute name must be in lowercase and should be followed by an equal sign and the desired value:

    % logger -e UI_COMMIT -a username=user -a command=commit

6. The syslog message follows all the command line arguments. Quotes are not required but are recommended for clarity:

    % logger -e UI_COMMIT -d mgd "This is a fake commit."

The above command causes the following message to be shown in the syslog:


    Jul 22 12:47:03 Junos mgd: UI_COMMIT: This is a fake commit.

NOTE When using the logger utility the event ID must always be in uppercase and the attribute names must always be in lowercase.


So let’s test this out.


root@vSRX-NAT-GW% logger -e LICENSE_VIOLATION -p "This is a test of the emergency broadcast system"
root@vSRX-NAT-GW% exit
root@vSRX-NAT-GW> show log messages | match LICENSE_VIOLATION | last 10
Jan 27 04:53:59 vSRX-NAT-GW logger: LICENSE_VIOLATION: This is a test of the emergency broadcast system


You can further confirm this by checking the syslog configuration on your local device. In this case, I’ve specified I’m ONLY sending messages to external host if the facility is ‘external’ AND the severity is ‘info’ or greater (ie. not debug) AND the regex of the message matches LICENSE. Otherwise, we’ll likely have a local catch-all configured with any-any to locally log messages we may not be explicitly interested in looking at on the remote server.

root@vSRX-NAT-GW> show configuration
## Last commit: 2016-01-27 04:59:04 UTC by root
version 12.1X47-D20.7;
system {
  syslog {
    host {
      external info;
      match LICENSE;
   file messages {
     any any;
     authorization info;
   file interactive-commands {
     interactive-commands any;


To check this is working correctly, we’ll throw a ‘monitor traffic’ on the management interface looking for outbound UDP traffic to my syslog host, and we’ll generate some messages.

root@vSRX-NAT-GW> monitor traffic interface ge-0/0/0 matching "host and udp"
Use <no-resolve> to avoid reverse lookups on IP addresses.
941 packets received by filter


And here’s the various tests we ran;

barnesry-mbp:python barnesry$ ssh root@
--- JUNOS 12.1X47-D20.7 built 2015-03-03 21:53:50 UTC
root@vSRX-NAT-GW% logger -e LICENSE_VIOLATION -p daemon.warning "This is a test"
root@vSRX-NAT-GW% logger -e VIOLATION -p "This is a test"
root@vSRX-NAT-GW% logger -e VIOLATION -p "This is a test"

You can see in the capture above we captured no packets leaving my local host as a result of any of the above tests we generated.

Now, let’s restart the packet capture again, and generate a message we’re pretty certain will match.

barnesry-mbp:python barnesry$ ssh root@
--- JUNOS 12.1X47-D20.7 built 2015-03-03 21:53:50 UTC
root@vSRX-NAT-GW% logger -e LICENSE_VIOLATION -p "This is a test"


Let’s check our packet capture again.

root@vSRX-NAT-GW> monitor traffic interface ge-0/0/0 matching "host and udp"
05:09:17.723090 Out IP truncated-ip - 42 bytes missing! > SYSLOG, length: 74
121 packets received by filter
0 packets dropped by kernel



Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s