Obtaining good thresholds for monitoring SFP light levels
We would like to alert if an SFP's received light level is too low. The light level sensor data is present and linked with a specific port so we can adjust the alert checkers based on attributes of the linked port. What we haven't figured out how to do is get appropriate thresholds for alerting for each SFP.
The SFPs define warning and alarm thresholds which are readable on the devices containing the SFPs but that data doesn't make it into Observium for most of our monitored devices as far as we can tell. Similar SFPs usually have similar thresholds so grouping by SFP vendor and part number then creating a check for each set of SFPs with manually extracted and entered thresholds could be acceptable. Observium does show the SFP vendor and part number on the device overview and inventory pages along with which port contains the SFP but there doesn't seem to be a way to reference this data when creating port or sensor groups or alert checkers.
There are also the automatically generated or manually set thresholds for sensors inside Observium itself. The automatically generated thresholds are often missing and we are not aware of a way to ensure they are present or reasonable. Manually set thresholds could work if there was a way to set and audit those thresholds in an automated fashion. We don't see a way to use the API to do this though (we're using the subscription edition).
Are there recommendations on how to achieve this? Is this a solved problem and we're running into a bug? This is feeling more difficult than it should be.
Corwin
participants (1)
-
Corwin Ziegler Hunts