I had origonally been under the impression that when an agent marked an item as "unsupported" it essentially disabled that item on the agent... at least until the next time it's restarted.
However, based on closer inspection that doesn't appear to be the case. Even if unsupported, it keeps trying and failing as unsupported over time as indicated by the agent logs.
Ultimately this is a performance impacting issue that should be resolved.
To illustrate my (flawed) understanding and how I am using it... I have several types of machines all under the same template. Some have igb network interfaces, some have e1000g interfaces. Rather than have to specify which is which I simply created network items for both types of interfaces with the expectation that the incorrect items will just go unsupported and disable themselves, thus saving me a lot of per-server hand holding.
A flag in the agent config would be nice to determine the appropriate behavior.
However, based on closer inspection that doesn't appear to be the case. Even if unsupported, it keeps trying and failing as unsupported over time as indicated by the agent logs.
Ultimately this is a performance impacting issue that should be resolved.
To illustrate my (flawed) understanding and how I am using it... I have several types of machines all under the same template. Some have igb network interfaces, some have e1000g interfaces. Rather than have to specify which is which I simply created network items for both types of interfaces with the expectation that the incorrect items will just go unsupported and disable themselves, thus saving me a lot of per-server hand holding.
A flag in the agent config would be nice to determine the appropriate behavior.