It looks like “Browser Instance Crashed” errors are not logged in the downloadable log files and are only visible in the web UI.
Regards,
Heinz
Hi Heinz,
Huh, that’s bizarre. The log lines that end up in the Browsertrix UI are pushed to Browsertrix via Redis by the crawler after being written to the crawl logs, so these lines should end up in both places. Maybe there’s a bug with these particular lines being written to the log stream - there is a try/catch block there that doesn’t log any details on error. I’ll try to see if I can reproduce the issue.
Tessa
Ah okay, I remember now. I think you’re seeing the one specific exception to the logging, where the Browsertrix backend’s operator logs when the crawler itself has significant crashes: browsertrix/backend/btrixcloud/operator/crawls.py at main · webrecorder/browsertrix · GitHub. This log message is pushed to the Browsertrix UI for user visibility into the state of their crawl, but is otherwise in the Browsertrix backend logs rather than the crawler logs.