Werk #22962: Parse messages that aren't valid UTF-8
| Component | Event Console | ||
| Title | Parse messages that aren't valid UTF-8 | ||
| Date | Nov 2, 2026 | ||
| Level | Trivial Change | ||
| Class | Bug Fix | ||
| Compatibility | Compatible - no manual interaction needed | ||
| Checkmk versions & editions |
|
The Event Console didn't parse a message if it contained even a single byte that isn't valid UTF-8, e.g. text in Latin-1, or a character cut in half by a sender truncating its messages. It created an event with the priority "emerg", the facility "user", neither host nor application, and the whole message as its text instead. So rules matching on host or application didn't match such events, and the events weren't associated with their hosts.
Now the Event Console replaces such bytes with the Unicode replacement character U+FFFD ("�") and parses the message as usual. This applies to all messages it receives as text, e.g. via syslog and from notifications forwarded with mkevent.