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
The idea is to improve the error messages outputted by Telegraf.
Proposal:
It would be very handy to have two additional information in the error message (when appropriate):
Server (SQL Server instance) - the instance in which the error occurred
Query - the query that was running (the name written in the config file, ie: Schedulers, PerfCounters)
Current behavior:
In most cases when you receive an error (and you are monitoring more than one instance) you are just blind about where the error happened.
Some examples of the actual behavior:
permission errors, the message is something like "login failed for user xxx" but SQL server itself does not tell you in which instance it happened
Query related error, in those cases you just see a section of the SQL query, of course, you can search in the code for that bit, but it can be generic, and anyway, you don't know in which instance it happened.
other errors already report the SQL instance who raised them (because the info is contained in the SQL Server error itself), like connection errors and point parsing error (if they show you the point who caused the error you can look at the "sql_instance" tag value)
The current error message look like this: _timestamp_ E! [input.sqlserver] Error in plugin: _errorMessage_
Desired message _timestamp_ E! [input.sqlserver] Error in plugin: Error on _SqlInstance_ while executing query _queryName_ : _errorMessage_
Desired behavior:
The error message includes the Sql instance who raised the error and the query (when appropriate)
Use case:
It will make error tracking way simpler when monitoring more than one instance
The text was updated successfully, but these errors were encountered:
I have addressed part of this aka the "Collector" part ( not the instance part) in #7934 . The instance part have to think about as didn't have instance name stored anywhere so has to be extracted from connections string. I didn's see this before made the changes. I agree adding instance at some point would be good too - specially for stuff like login failed :) shich is painful to figure out if have a bunch of connection strings.
This is what would look like
[input.sqlserver] Error in plugin: CollectorName : errorMessage
This has been partially fixed also by the query formatting. In case of errors relate to the SQL version an error is returned, which contains also the Instance name
Feature Request
The idea is to improve the error messages outputted by Telegraf.
Proposal:
It would be very handy to have two additional information in the error message (when appropriate):
Current behavior:
In most cases when you receive an error (and you are monitoring more than one instance) you are just blind about where the error happened.
Some examples of the actual behavior:
other errors already report the SQL instance who raised them (because the info is contained in the SQL Server error itself), like connection errors and point parsing error (if they show you the point who caused the error you can look at the "sql_instance" tag value)
The current error message look like this:
_timestamp_ E! [input.sqlserver] Error in plugin: _errorMessage_
Desired message
_timestamp_ E! [input.sqlserver] Error in plugin: Error on _SqlInstance_ while executing query _queryName_ : _errorMessage_
Desired behavior:
The error message includes the Sql instance who raised the error and the query (when appropriate)
Use case:
It will make error tracking way simpler when monitoring more than one instance
The text was updated successfully, but these errors were encountered: