Re: false positive http probe reports from routine 408
Gilles Detillieux <grdetil-b3I9/Z444iDkHagxZOarhfd9D2ou9A/[email protected]> Tue, 2 Oct 2018 12:25:10 -0500 (CDT)
| Newsgroups | gmane.comp.log.logwatch.devel |
|---|---|
| Message-ID | <[email protected]> |
Earlier, I wrote:
> There's a long-standing false-positive in the http service script which
> I finally got around to looking into and patching. When Apache logs 408
> Request timeout errors, it logs these with a null request field (just a
> "-"). These are routine errors when the client keeps the socket open too
> long without a request, but they're reported as probes/known hack
> attempts. This patch quiets down that noise, just giving a count of 408
> errors at the end...
Unfortunately Thunderbird changed a bunch of the spaces in the patch to
non-breaking spaces, which might make the patch hard to apply unless you
hand-edit. Here's a cleaner version...
--- logwatch-7.4.3/scripts/services/http.orig 2016-03-30 16:32:34.000000000 -0500
+++ logwatch-7.4.3/scripts/services/http 2018-10-02 10:55:19.000000000 -0500
@@ -387,6 +387,8 @@ while (my $line = <STDIN>) {
($field{method},$field{url},$field{protocol}) = split(/
+/,$field{"request"});
if (! $field{url}) {
$field{url}='null';
+# don't count timeout errors as null requests...
+ $field{url}='-' if $field{http_rc} == 408;
}
$field{lc_url} = lc $field{url};
--
Gilles R. Detillieux E-mail: <grdetil-b3I9/Z444iDkHagxZOarhfd9D2ou9A/[email protected]>
Spinal Cord Research Centre WWW: http://www.scrc.umanitoba.ca/
Dept. of Physiology and Pathophysiology, Rady Faculty of Health Sciences,
Univ. of Manitoba Winnipeg, MB R3E 0J9 (Canada)