When a parser fails to parse incoming input, LogScale automatically adds the following fields to the event:
@error_msg: contains the error message
You can search for these fields to determine what happened and update the parser accordingly. When you find an error, you can add the @rawstring of any events that fail to parse as a test case for your parser.
You can search for all events that were not parsed correctly by searching the repository:
You can extend the query in different ways. For example, you can display a timechart:
@error=* | timechart()
or group results by error message:
@error=* | groupBy(@error_msg)
The original log input is available in @rawstring.
There is no way for LogScale to re-parse data once it's stored, even if there is an error on the event. You'll have to resend the event through the ingest API.