You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Interrupt resolution to cause an error, such as a timeout.
Expected behavior:
One of:
query_time_ms=10000
null value
Actual behavior:
query_time_ms=0
Additional info:
This is highly problematic as the value indicates that the response came back immediately. However this is not the case. The response didn't come back at all. The result should somehow indicate this by either setting the value to the maximum (the timeout value), or not setting it at all, and providing some other field indicating an error occurred.
The text was updated successfully, but these errors were encountered:
I think it makes sense to have a field that is emitted only when a timeout occurs with the value being the time waited: query_timeout_ms=10000. I wonder if we should try to stick to the same units across fields, should timeouts always be reported in seconds? Certainly this would be silly for some points so maybe not.
Bug report
Relevant telegraf.conf:
System info:
Telegraf 1.2.1
Steps to reproduce:
Expected behavior:
One of:
query_time_ms=10000
Actual behavior:
query_time_ms=0
Additional info:
This is highly problematic as the value indicates that the response came back immediately. However this is not the case. The response didn't come back at all. The result should somehow indicate this by either setting the value to the maximum (the timeout value), or not setting it at all, and providing some other field indicating an error occurred.
The text was updated successfully, but these errors were encountered: