使用logstash解析包含python回溯的日志

Question

我一直在尝试使用logstash解析python回溯日志。我的日志如下：

[pid: 26422|app: 0|req: 73/73] 192.168.1.1 () {34 vars in 592 bytes} [Wed Feb 18 13:35:55 2015] GET /data => generated 2538923 bytes in 4078 msecs (HTTP/1.1 200) 2 headers in 85 bytes (1 switches on core 0)
Traceback (most recent call last):
  File "/var/www/analytics/parser.py", line 257, in parselogfile
    parselogline(basedir, lne)
  File "/var/www/analytics/parser.py", line 157, in parselogline
    pval = understandpost(parts[3])
  File "/var/www/analytics/parser.py", line 98, in understandpost
    val = json.loads(dct["events"])
  File "/usr/lib/python2.7/json/__init__.py", line 338, in loads
    return _default_decoder.decode(s)
  File "/usr/lib/python2.7/json/decoder.py", line 366, in decode
    obj, end = self.raw_decode(s, idx=_w(s, 0).end())
  File "/usr/lib/python2.7/json/decoder.py", line 382, in raw_decode
    obj, end = self.scan_once(s, idx)
ValueError: Unterminated string starting at: line 1 column 355 (char 354)

到目前为止，除了最后一行之外，我已经能够解析日志

^{pr2}$

我使用多行过滤器来实现这一点。我的logstash配置如下所示：

filter {

    multiline {
        pattern => "^Traceback"
        what => "previous"
    }

    multiline {
        pattern => "^ "
        what => "previous"
    }


    grok {
        match => [
            "message", "\[pid\: %{NUMBER:process_id:int}\|app: 0\|req: %{NUMBER}/%{NUMBER}\] %{IPORHOST:clientip} \(\) \{%{NUMBER:vars:int} vars in %{NUMBER:bytes:int} bytes\} \[%{GREEDYDATA:timestamp}\] %{WORD:method} /%{GREEDYDATA:referrer} \=\> generated %{NUMBER:generated_bytes:int} bytes in %{NUMBER} msecs \(HTTP/%{NUMBER} %{NUMBER:status_code:int}\) %{NUMBER:headers:int} headers in %{NUMBER:header_bytes:int} bytes \(%{NUMBER:switches:int} switches on core %{NUMBER:core:int}\)%{GREEDYDATA:traceback}"
            ]
    }

    if "_grokparsefailure" in [tags] {
        multiline {
            pattern => "^.*$"
            what => "previous"
            negate => "true"
        }
    }

    if "_grokparsefailure" in [tags] {
        grok {
            match => [
                  "message", "\[pid\: %{NUMBER:process_id:int}\|app: 0\|req: %{NUMBER}/%{NUMBER}\] %{IPORHOST:clientip} \(\) \{%{NUMBER:vars:int} vars in %{NUMBER:bytes:int} bytes\} \[%{GREEDYDATA:timestamp}\] %{WORD:method} /%{GREEDYDATA:referrer} \=\> generated %{NUMBER:generated_bytes:int} bytes in %{NUMBER} msecs \(HTTP/%{NUMBER} %{NUMBER:status_code:int}\) %{NUMBER:headers:int} headers in %{NUMBER:header_bytes:int} bytes \(%{NUMBER:switches:int} switches on core %{NUMBER:core:int}\)%{GREEDYDATA:traceback}"
        ]
            remove_tag => ["_grokparsefailure"]
        }
    }
}

但我的最后一行不是解析。相反，它仍然会给我一个错误，并且会成倍地增加处理时间。关于如何解析回溯的最后一行有什么建议吗？在

Answer 1

好吧，我找到了解决办法。因此，我采用的方法是忽略以'['开头的日志消息的开头，所有其他行都将附加到前一条消息的末尾。然后可以应用grok过滤器并解析回溯。请注意，我必须应用两个grok过滤器：

1. 当有GREEDYDATA的回溯时。

2. 因为在没有回溯的情况下，GREEDYDATA解析会失败，我将不得不删除grokparsefailure标记，然后在没有GREEDYDATA的情况下再次应用grok。这是在if块的帮助下完成的。

最后的logstash过滤器如下所示：

filter {

    multiline {
        pattern => "^[^\[]"
        what => "previous"
    }



    grok {
        match => [
            "message", "\[pid\: %{NUMBER:process_id:int}\|app: 0\|req: %{NUMBER}/%{NUMBER}\] %{IPORHOST:clientip} \(\) \{%{NUMBER:vars:int} vars in %{NUMBER:bytes:int} bytes\} \[%{GREEDYDATA:timestamp}\] %{WORD:method} /%{GREEDYDATA:referrer} \=\> generated %{NUMBER:generated_bytes:int} bytes in %{NUMBER} msecs \(HTTP/%{NUMBER} %{NUMBER:status_code:int}\) %{NUMBER:headers:int} headers in %{NUMBER:header_bytes:int} bytes \(%{NUMBER:switches:int} switches on core %{NUMBER:core:int}\)%{GREEDYDATA:traceback}"
        ]
    }

    if "_grokparsefailure" in [tags] {
        grok {
            match => [
            "message", "\[pid\: %{NUMBER:process_id:int}\|app: 0\|req: %{NUMBER}/%{NUMBER}\] %{IPORHOST:clientip} \(\) \{%{NUMBER:vars:int} vars in %{NUMBER:bytes:int} bytes\} \[%{GREEDYDATA:timestamp}\] %{WORD:method} /%{GREEDYDATA:referrer} \=\> generated %{NUMBER:generated_bytes:int} bytes in %{NUMBER} msecs \(HTTP/%{NUMBER} %{NUMBER:status_code:int}\) %{NUMBER:headers:int} headers in %{NUMBER:header_bytes:int} bytes \(%{NUMBER:switches:int} switches on core %{NUMBER:core:int}\)"
                ]
            remove_tag => ["_grokparsefailure"]
        }
    }

    else {
        mutate {
            convert => {"traceback" => "string"}
        }
    }

    date {
        match => ["timestamp", "dd/MM/YYYY:HH:MM:ss Z"]
        locale => en
    }
    geoip {
        source => "clientip"
    }
    useragent {
        source => "agent"
        target => "Useragent"
    }
}

或者，如果您不想使用if块检查另一个grok模式并删除_grokparsefailure，那么可以使用第一个grok过滤器来检查这两种消息类型，方法是在grok filter的match数组中包含多个消息模式检查。可以这样做：

^{pr2}$

还有第三种方法（可能是最优雅的一种）。看起来像这样：

grok {
    match => [
        "message", "\[pid\: %{NUMBER:process_id:int}\|app: 0\|req: %{NUMBER}/%{NUMBER}\] %{IPORHOST:clientip} \(\) \{%{NUMBER:vars:int} vars in %{NUMBER:bytes:int} bytes\} \[%{GREEDYDATA:timestamp}\] %{WORD:method} /%{GREEDYDATA:referrer} \=\> generated %{NUMBER:generated_bytes:int} bytes in %{NUMBER} msecs \(HTTP/%{NUMBER} %{NUMBER:status_code:int}\) %{NUMBER:headers:int} headers in %{NUMBER:header_bytes:int} bytes \(%{NUMBER:switches:int} switches on core %{NUMBER:core:int}\)(%{GREEDYDATA:traceback})?"
    ]
}

注意，在这个方法中，存在为可选的字段必须用“（）？”括起来。这里，(%{GREEDYDATA:traceback})?

因此，grok过滤器会看到如果字段可用，它将被解析。否则，将跳过它。在