Closed St3f1n closed 11 months ago
However, the Inputs.internal shall not raise an error from the ping plugin because internal is independent from the network.
The errors reported by the internal plugin are any and all errors by a plugins. It does not differentiate between network errors, retryable errors, or any other types. This is not something we would change at this time as it would require specifying different types of errors across the entire codebase.
My point of view is that when the ping plugin receives error return 2, this should not get treated as an error in general, rather a normal result. Secondly especially for the internal_gather states.
Relevant telegraf.conf
Logs from Telegraf
System info
Docker
-
Steps to reproduce
Running telegraf with and without network connection.
Expected behavior
The Inputs.ping shall report the ping error = 2 as is. However, the Inputs.internal shall not raise an error from the ping plugin because internal is independent from the network.
The Input.ping error_return 2 should not get treated as an error in general, rather as a normal result.
Actual behavior
I'd like to use the Inputs.internal for alerting purposes if some "internal" error happen and i want to have them independent from the network stability. Unfortunately at the moment i get those "internal" errors when the network is unstable and therefore i need to separate the ping plagin completely from overlaid alerting process.
Line 7 provides the expected return_code 2. Line 8, 12, ... then presents errors=1 (which i'd expect to stay 0). Note that lines 2 and 5 sent errors=0.
outfile:
Additional info