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
{{ message }}
Repository navigation
exception happened in background worker: source sequence is illegal/malformed utf-8 #2805
Something fails in production and all I see in the logs is this message,
which is not helpful since I need the original backtrace and error message to begin debugging
Reproduction Steps
raise an error that includes invalid utf8
Expected Behavior
<?> icons replace invalid utf8
Actual Behavior
error not sent to sentry and log shows message
Ruby Version
3.4.6
SDK Version
5.28.0
Integration and Its Version
none
Sentry Config
nothing special, except a custom logger that does {message: message}.to_json
@grosser thanks for the report. Any chance you could give me an example message that causes this? We do have handling of malformed utf-8 strings in place.
sadly did not find more info and don't have a way to reliably reproduce it
feel free to close if you don't see any obvious spot where this could have slipped through
Activity
linear commentedon Dec 2, 2025
RUBY-123
solnic commentedon Dec 9, 2025
@grosser thanks for the report. Any chance you could give me an example message that causes this? We do have handling of malformed utf-8 strings in place.
grosser commentedon Dec 9, 2025
sadly did not find more info and don't have a way to reliably reproduce it
feel free to close if you don't see any obvious spot where this could have slipped through
solnic commentedon Dec 10, 2025
@grosser alright, I'll dig into it - could you tell me what exact custom logger do you use that does
{message: message}.to_json?20 remaining items
dry & add sanitized original msg in stderr output
dingsdax commentedon Jan 5, 2026
@grosser fyi, we decided for a slightly different approach, guarding against any errors in custom loggers and output to stderr in case #2817
grosser commentedon Jan 5, 2026
yeah that looks like the right approach, there could be various errors with logging and this kinda solves them all
fix: prevent SDK crash when SDK logging fails (#2817)