For the UDP input metrics, make it clear when the data is invalid. The availability of these two metrics depends on the host:
system_packet_drops
receive_queue_length
My suggestion is to report the values as -1 at initialization or after any error.
Describe a specific use case for the enhancement or feature:
A system_packet_drops value of 0 is a indicator of health, but if zero is being reported because of a collection failure (or because we are on Windows), then we have a false sense of the current state. When troubleshooting I need to know if the metric is valid.
Describe the enhancement:
For the UDP input metrics, make it clear when the data is invalid. The availability of these two metrics depends on the host:
system_packet_drops
receive_queue_length
My suggestion is to report the values as
-1
at initialization or after any error.Describe a specific use case for the enhancement or feature:
A
system_packet_drops
value of0
is a indicator of health, but if zero is being reported because of a collection failure (or because we are on Windows), then we have a false sense of the current state. When troubleshooting I need to know if the metric is valid.References