Default health sensors

I wrote something about this topic yesterday. For my cisco routers the defaults pulled from SNMP are ignored and replaced with librenms generated defaults that are far too close to the sensor readings and keep generating spurious alerts.

Still trying to understand if this is a bug that other people are seeing or if there’s something up with my installation. Trying to work my way through whats happening when the discovery is done.

juniper, ubnt, hp, apc - the temperature threshholds are completely off, but i preferre not to change it manually per device

sound a lot like what i’m working on

that’s very odd I running APC and HP devices and haven’t had to change any of the default sensor values.

like I said, guys if there is no support from the vendor’s mib… then libreNMS will just auto calculate the values.

So at that point, you will need to customize the values in health settings for that device.

this has been discussed b4 see here Cisco ONS sensor under limit alerts as of 08/14

its mostly the devices placed outdoor, and the climate here is a little colder than what the vendors expect.

when the sun goes down i get about 8-10 alarms about sensor under limit, but the devices are well within vendor specifications, så the thresholds are too narrow

URL to your post?

Yes, so you will need to adjust to health sensor high low values for the device in the web UI.
I had do that on a few devices also that are in odd environments.

Right, I get that, but the MIB has the threshold values and it looks like the scripts (I’m looking at cisco-entity-sensor.inc.php and friends) are polling the right MIB and oids, the value just aren’t making into mysql.

I was really wondering if it was just me that was seeing this. I’ll try and dig further and provide more info as and when I find it.

if find something that needs to be changed in the code, submit it in git hub.

well i can see that the high low varies from several devices, same vendor, type, os, fw, so i guess they are calculated.

oh well, i’ll then do the changes manually, but would have been nice with a feature to overrule the values per device/vendor

John, take a look at /opt/librenms/includes/discovery/functions.inc.php, that has the routines where the values are calculated if they can’t be extracted from the MIB.

i will look into that - so far ive set the values static

Hey All,

i have quiet the same issue - The values for various devices are totally off.
For example my QNap NAS in a small branch office:


Every now and then (once per week?) the values are reset. I edit the device and lower the values of the fan sensors, a week later they are set back to a ways to high value. Result: I’m getting alarms for too low fan sensor values (which are still very healthy).

any advice?

Hay All

We have similar situation. LibreNMS reported that Cisco device temperature sensor over limit.
Found that there is incorrect values on Health settings. Corrected manually but next day wrong values back again.

Hi

Just want to check if you been able resolve mentioned issue?

Hi - I’m facing same issue - looks like you also need to set in MySQL sensors table the column sensor_custom to Yes then it shouldn’t overwrite them - as if its set to No it seems to re-read/guess them at certain points