Alerting off of Error DB

[Jason Russell]

Is there anyone running alerts on the Error DB (mainly new additions to it)? I’m wanting to shut down a thread based on errors being entered into the Error DB for those threads. However, we’ve not been able to get the alert to trigger, whether it’s to shut down, email, post an alert to GM, nothing. Has anyone done this? Is there some setting that makes it work more/less than others?

Settings (Mind you, we’ve fiddled with these a LOT)

Alert Type: Error Database

Source: (A single thread, all threads, full list of threads have been tried)

Source Count: Any, All, At Least

Comparing: > 0, >=1, >1

Duration: Once

Repeating: every 1 second (and off, and minute)

Time Window – Set various specific  times to check.

 

Is there something we’re missing/doing wrong? Just can’t seem to get it to trigger.

Topic

[Jay Hammond]

We have an email alert set up for the error DB in all sites that will send an email to us if there are any entries in it. We don’t have any commands to shut down the threads, but here’s our config:

 

Alert Type: error database

Source: highlighted all threads

Comparing: > 0

Duration: once

Repeating:  N minutes 90 Max: 2

Time Window: */5:30-17:30/*/*

 

It’s worked well for us for several years now. You might want to check with your email admin that the user is allowed to send emails in your network. That bit us a few times.

[Jason Russell]

Yeah, we had to mess with the alerts and emails a bit too, but we did get emails to pass through (one is set to interface down and it works fine).

What is your source count set to?

Also, looking at your “time window” , we were setting for a single time, and i see now your s looks like from 5:30-17:30 on the hour slot. That may be part of our issue.

[Jay Hammond]

Oh, sorry I left that one out.  Source Count is set to Any.

[Jason Russell]

Amusingly enough, now it just crashes the monitorD if I get an error:

{ALERT
{ NAME {Error DB Depth – st_sched1 Test} }
{ VALUE errdb }
{ SOURCE t_eGSCH2aria_o }
{ WITH -2 }
{ COMP {> 0} }
{ FOR once }
{ REPEATING {
{ MAX 2 }
{ TIME {nsec 1} }
} }
{ WINDOW */5:30-17:30/*/* }
{ CUSTOMMSG {{ITEM {}} {DELIMITER keyvalue}} }
{ ACTION {
{ email {
{ FROM cloverleaf@infor.com }
{ TO jmcrussell@firsthealth.org }
{ SUBJ {Error DB Test} }
{ MSG {Test to trigger. } }
} }
} }
}

 

Everything runs fine until an error hits, and then everything goes down.  The monitor immediately crashes after I attempt to start it. Once I clear the errors, it works fine. Any ideas?

[Jay Hammond]

The only thing that jumps out at me is that the thread name contains upper case letters. There’s a Best Practice document that states:

<td width=”107″>Naming

Convention
<td width=”107″>Site, Process, and Threads
<td width=”292″>Always use lowercase letters for site, process and thread names.
<td width=”167″>Not required for other CL objects

But, we have threads with upper case names and it doesn’t bomb MonitorD.

[Jason Russell]

Finally decided to contact support about this, and after a lot of testing, and getting close to going full debug mode, we found the issue. It seems that cloverleaf writes out a temporary mail file in $HCISITE/exec/hcimontord/. However, something happened with the files we wrote out, and cloverleaf wasn’t able to overwrite these files or change them. Once the files were removed, the alert worked fine. Something very odd and not common at all that was at play.

[Author]

Replies