Monitor Dell Server for Temperature, Power Supply and Other properties

Hello nberglin,
thank you for your KB-post.

• Are you querying the Server via the iDRAC interface or the Operating System's address?

On the sensor's manual page there's a screenshot showing "usual" sensor's results:

How exactly does your sensor look like? Please share a screenshot.

Best Regards,
Luciano Lingnau [Paessler Support]

Hi,

After reading some articles from this site and others, in order for PRTG to talk to our equipment we had to install Dell OpenManage Server Administrator.

It appears I am querying the server via IP address.

When we use the sensor we get the same exact results as your screen shot above (see link for our screenshot). http://i.imgur.com/okm0KYn.png

Yes, if you want to monitor the Server by it's OS Address you must install the Openmanage software which will "extend" the SNMP support for hardware-specific MIB's. When querying it via the iDRAC this is not required.

Please let us know if you require any further assistance.

Best Regards,
Luciano Lingnau [Paessler Support]

I get that I needed to install OMSA on my machine which was done prior to this thread. When I added the SNMP Dell PowerEdge System Health sensor, PRTG talks to it fine. when we enact a non-critical error we get expected results (example, we set allowed temperature threshold to x and get PRTG to tell us somethings not right which is expected: http://i.imgur.com/WhFyZjg.jpg).

Half way to my goal however. What I want is to get PRTG to tell us more then: "Normal, non-critical, critical" status, but usable actionable, data.

I tried added a SNMP Dell Hardware sensor and added the appropriate fields for what I want to track, but instead of getting a nice data value like internal temp is 42 C, I get 210 or some other non specified value. http://i.imgur.com/L23wJ5p.png

Now I've read in places (including your last reply) that since OMSA is installed, I need to then use the hardware specific MIB values and import/use them in PRTG. Understandable. But then I run into what this fine person ran into: https://kb.paessler.com/en/topic/47063-dell-2950-+-custom-mibs

Different OID values from different dell specific MIB libraries (http://www.dell.com/support/home/us/en/19/Drivers/DriversDetails?driverId=HR34Y) still result to either an error 222 or 223.

At this point, I'm a bit confused and I am unsure of what to try next. Maybe I need a new sensor or tweak something different?

Hello, thank you very much for the detailed reply, I wasn't aware that you were interested in the device's readings(instead of state), that makes it much clearer.

If you want to monitor the actual values, you'll have to locate them in the vendor's MIB to know "where to look for". I've checked the MIB-Dell-10892 and located the following:

Those OID's are table which should contain multiple rows and entries, one for each "entity" of that type.

Please download our latest SNMP Tester and perform a "Walk" against the target device providing the OID's above one at a time and run the test, some may not return anything based on the information available for your hardware but please run the test for all these OID's.

Please share the results of the test above.

On a side note, please be aware that conceptually it makes a lot more sense to monitor for "state" and not "value". PRTG includes several types of alerts and notifications, you don't need to constantly review your monitoring data - With "states/status" PRTG can let you know exactly when something requires your attention. I'm not saying that you shouldn't do this, just that you don't have to(but I'm glad to assist you with it because these concepts also apply to monitoring devices for which we don't have native sensors).

After I posted this reply, I ended up using that tool again (after some self education on what it really does, was mistaken that it did something else).

I ended up getting the data values that I wanted. Now that I see what the raw data value is coming through (through the SNMP tester/walker) I adjusted the sensor I added earlier (http://i.imgur.com/L23wJ5p.png) which makes sense: OID value says 200, the sensor says 200. Cool.

I suppose that there is not a way to tell the sensor to chop of an extra '0' or add a decimal point so it reads '20°' or '20.0°' is there? I tried using the delta counter instead of a gauge when added the value but got unexpected/unwanted results. It doesn't seem like there is, but I figured I'd ask before assuming.

Hello, thank you for your post.

If you're using a recent PRTG version, you should be able to use a multiplication/division factor in the channel's settings, please review the following: SNMP Advanced Division/Multiplication

Dividing by 10 will get you 20.0, this is a very common approach for reporting temperature via SNMP.

Alternatively, if you switch to the SNMP Custom sensor you will be able to enter this in the sensor's settings.

Best Regards,
Luciano Lingnau [Paessler Support]

Again, this ended up working. The original reason why I asked was because I was getting this: http://i.imgur.com/VTxFh8i.png

I found I was using the Delta Counter then dividing that number because I was totally oblivious to the fact that under 'Gauge (unsigned integer)' was the option to multiply/divide.

It is now working, reading at 19°C atm. Appreciate the help!