by Nevin1901 on 10/21/23, 8:56 PM with 1 comments
by avidiax on 10/22/23, 1:47 AM
* Logging in a large system is a tragedy of the commons. A few players spam the logs like crazy, and the rest often don't log enough. Logs, logs, everywhere, but not a line to debug.
* Logging tools do a very bad job revealing discrepancies in timing. i.e. did the gap between two events change?
* Very few tools are able to continuously ingest logs and then highlight discrepancies or differences in a log of interest. That could be lines that are different, have increased in frequency, or are absent. I have sed scripts to be able to compare good and bad logs in this way, but this is a last resort when nothing else helps.
*