I have an SNMP trap receiver sensor set up at the top of my device tree, on the local probe device. It works as expected and receives SNMP traps from any device in the tree, which is what we want. However, the error message that is generated is generic, either the default

"0.01 #/s (Errors) is above the error limit of < 0.01 #/s"

or a fixed message that I create. Is there any way to insert OIDs into the notification message so that it's obvious what device it's coming from and/or what the trap is?

If I replaced the single sensor with one sensor for each device, that could show me the origin of the alert, but it still wouldn't identify the trap (and it would be a bit of a pain to set up).

Also, a second question about the SNMP trap sensor:

The way the trap receiver sensor works is that the error clears and goes green as soon as the next polling cycle happens, i.e. within a minute. Is there any way to keep the alert red until it's specifically cleared? For example, if we have a failed disk drive in an array somewhere, it's helpful to get alerted that the drive needs replacing, but it would be more helpful to have the alert remain red until the drive is actually replaced and the error clears.

Article Comments

Dear David

The OID cannot be put into the status message, or notification message. However, you can use these related placeholders for the notification message:

  • %trapmessages
  • %trapwarnings
  • %traperrors

Please find all available placeholders here.

The sensor only shows the bandwidth of traps, not the total accumulated amount. You can use the sensor to trigger a notification, but the sensor status itself will clear with the next scan.


Aug, 2015 - Permalink

Thanks. That helps. It's too bad I can't put anything in the notification subject, but for now putting %trapmessages in the message body will have to do.


Aug, 2015 - Permalink