From 9bf5591ecdb9acb654d04a381c8d5fc9f8961a5d Mon Sep 17 00:00:00 2001
From: Valentin Gologuzov <vgologuz@redhat.com>
Date: Mon, 15 Jun 2015 14:13:35 +0200
Subject: [PATCH] [copr][backend] logststash config: ignore requests generated
 by search engine crawlers

---
 .../backend/templates/logstash.d/copr_backend.conf     | 10 ++++++++++
 1 file changed, 10 insertions(+)

diff --git a/roles/copr/backend/templates/logstash.d/copr_backend.conf b/roles/copr/backend/templates/logstash.d/copr_backend.conf
index 532dad34ba..6317342f18 100644
--- a/roles/copr/backend/templates/logstash.d/copr_backend.conf
+++ b/roles/copr/backend/templates/logstash.d/copr_backend.conf
@@ -18,6 +18,16 @@ filter {
       patterns_dir => "/usr/share/logstash/patterns"
       pattern => "%{LIGHTTPD}"
     }
+
+    # ignore requests generated by search engine crawlers
+    useragent {
+      source => "agent"
+      target => "ua_parsed"
+    }
+    if [ua_parsed][device] == "Spider" {
+      drop {}
+    }
+
     date {
        match => [ "timestamp", "dd/MMM/yyyy:HH:mm:ss Z" ]
     }