Node down alarm outside of polling time

[user54119]

For me the question is, what triggers the SYS_NODE_DOWN event. The default DCI 'status' has no threshold configured.

[alk]

It's generated internally, when server can't reach node via any allowed channel (ICMP, Agent, SNMP)

Easiest way is to have a rule in a event processing policy which silently ignores SYS_NODE_DOWN during off-hours.
Check attached screenshot for sample configuration.

Please note that this EPP is processed from top to bottom and by default all rules are processed (unlike "first-match-win" in thresholds), so this rule should be above any other SYS_NODE_DOWN related rules.

Description of "TIME" structure (returned by localtime()): http://wiki.netxms.org/wiki/NXSL:TIME

Update: script should be like "if (now->hour >= 20 || now->hour <= ..."

@user54119 said in Node down alarm outside of polling time:

For me the question is, what triggers the SYS_NODE_DOWN event. The default DCI 'status' has no threshold configured.

---

[victor]

Actually, there is an etrror in the script: you should use || (logical or) instead of && (logical and).

Best regards,
Victor

[user54119]

thanks for this 'workaround'. But in my opinion this is not very elegant, it's inconsistent. All DCIs of a node are managed by there own 'custom schedule', only for 'status' we need an event processing policy. A 'custom schedule' for the status of a node would be a good feature. Or maybe SYS_NODE_DOWN event could be triggered by the 'status' DCI, so the 'custom schedule' of this DCI could be used.

[victor]

It is possible to add threshold to status DCI and generate some custom event instead of SYS_NODE_DOWN, and do all processing for this custom event. But status and SYS_NODE_DOWN have one important difference - status also takes current alarms into consideration. So, if there is active critical alarm, status will be 4 (critical), but SYS_NODE_DOWN will not be generated, because node itself is reachable. Starting from upcoming 1.2.12 release, it is also possible to create DCI which will check node down flag. If you put threshold on such DCI and will generate some custom node down event, you'll get exactly what you need. To create such DCI, create new DCI with source "Internal" and parameter name "Dummy", and use the following transformation script:

return ($node->runtimeFlags & 0x04) ? 1 : 0;

This DCI will return 1 if node is down and 0 otherwise.

Best regards,
Victor

[jdowney]

Hi looking at doing the same we have some nodes which are either rebooted nightly or weekly. Is there a way to implement per node of no status updates during this time?

Thanks

[jdowney]

would it be an option to create a maintenance schedule for this?

I've tried to use Cron Maker to set a cron schedule within the task but it does not seem run. Any ideas on what I'm doing wrong or if there is a better way to do this. 

www.cronmaker.com -

0 30 23 ? * MON,TUE,WED,THU,FRI *

 (Every weekday - Monday - Friday 23:30)

Thanks

[jdowney]

Nudge, Nudge Anyone know what I'm doing wrong or is there an issue with 2.1-M2 which is why this is not working?

[victor]

Hi,

as far as I can tell your schedule is not valid cron expression. As stated on web site "Generated expressions are based on Quartz cron format". NetXMS supports subset of UNIX cron format (see here for example: https://en.wikipedia.org/wiki/Cron#Modern_versions). It only supports *, /, -, L, and lists. Try to change your schedule to

30 23 * * 1-5

Best regards,
Victor

[jdowney]

Thanks a lot this has resolved our issue with overnight daily reboots.
 
John