Comments on: Which log levels to use when? http://www.databasesandlife.com/which-log-levels-to-use-when/ Adrian Smith's blog Mon, 06 Feb 2012 05:20:25 +0000 hourly 1 http://wordpress.org/?v=3.3.1 By: Robin Salih http://www.databasesandlife.com/which-log-levels-to-use-when/comment-page-1/#comment-89 Robin Salih Tue, 05 Feb 2008 20:03:57 +0000 http://www.databasesandlife.com/which-log-levels-to-use-when/#comment-89 Yeah, that's basically how I do logging, although instead of TRACE I have DEBUG and SPAM. If you set SPAM on it will basically fill up a hard disc very quickly as it tends to log each sql statement, messages between servers, function calls etc. DEBUG is useful during the development stage, logs in more detail than INFO so you can basically follow the path of execution. Additionally there is CRITICAL, which is where you cannot recover from an error. Yeah, that’s basically how I do logging, although instead of TRACE I have DEBUG and SPAM. If you set SPAM on it will basically fill up a hard disc very quickly as it tends to log each sql statement, messages between servers, function calls etc. DEBUG is useful during the development stage, logs in more detail than INFO so you can basically follow the path of execution.

Additionally there is CRITICAL, which is where you cannot recover from an error.

]]> By: Toby http://www.databasesandlife.com/which-log-levels-to-use-when/comment-page-1/#comment-85 Toby Tue, 05 Feb 2008 14:10:24 +0000 http://www.databasesandlife.com/which-log-levels-to-use-when/#comment-85 I would put FATAL errors on a separate level. FATAL errors are (for me) errors that have resulted in a catastrophic failure of some component of the system. I work on a system that has multiple failover components and without the separate TAG of a fatal error we wouldn't know immediately that a component had died and been failed over. The FATAL errors trigger a sound event on a number of PCs around the desk so we can respond rapidly. (One of the sound effects is the red dwarf 'This ol' baby's crashed more times than a ZX81!', so you get the picture on how reliable the system is ;-)). I would put FATAL errors on a separate level. FATAL errors are (for me) errors that have resulted in a catastrophic failure of some component of the system. I work on a system that has multiple failover components and without the separate TAG of a fatal error we wouldn’t know immediately that a component had died and been failed over. The FATAL errors trigger a sound event on a number of PCs around the desk so we can respond rapidly. (One of the sound effects is the red dwarf ‘This ol’ baby’s crashed more times than a ZX81!’, so you get the picture on how reliable the system is ;-)).

]]>