Syslog is an industry-standard protocol for recording device logs, commonly used in network management, security management, and log auditing. You can use the GROK function in Simple Log Service (SLS) Domain Specific Language (DSL) to efficiently parse Syslog logs in different formats.
Overview
Syslog is widely used for system logs on Unix-like operating systems. Syslog messages can be recorded in local files or sent over the network to a receiving server. The server can centrally store Syslog messages from multiple devices or parse their content for further processing.
Problem
For a long time, no standard existed to regulate the Syslog format, resulting in inconsistent formatting or no defined format at all. This makes Syslog messages difficult to parse programmatically, as they can often only be treated as plain strings. Understanding how to parse Syslog logs in various formats is therefore essential.
Video tutorial
Introduction to Syslog protocol standards
Two common Syslog protocols are used in the industry: RFC 5424 (published in 2009) and RFC 3164 (published in 2001). The following sections describe the differences between these two protocols. Understanding these differences helps you use GROK to flexibly parse Syslog logs.
-
RFC 5424 protocol
The RFC 5424 protocol includes the following fields. For more information, see the official protocol document.
PRI VERSION SP TIMESTAMP SP HOSTNAME SP APP-NAME SP PROCID SP MSGIDThe following examples describe these fields:
""" Example1: <34>1 2019-07-11T22:14:15.003Z aliyun.example.com ali - ID47 - BOM'su user' failed for lonvick on /dev/pts/8 """ PRI -- 34 VERSION -- 1 TIMESTAMP -- 2019-07-11T22:14:15.003Z HOSTNAME -- aliyun.example.com APP-NAME -- ali PROCID -- None MSGID -- ID47 MESSAGE -- 'su user' failed for lonvick on /dev/pts/8 """ Example2: <165>1 2019-07-11T22:14:15.000003-07:00 192.0.2.1 myproc 8710 - - %% It's time to make the do-nuts. """ PRI -- 165 VERSION -- 1 TIMESTAMP -- 2019-07-11T05:14:15.000003-07:00 HOSTNAME -- 192.0.2.1 APP-NAME -- myproc PROCID -- 8710 STRUCTURED-DATA -- "-" MSGID -- "-" MESSAGE -- "%% It's time to make the do-nuts." """ Example3: - with STRUCTURED-DATA <165>1 2019-07-11T22:14:15.003Z aliyun.example.com evntslog - ID47 [exampleSDID@32473 iut="3" eventSource= "Application" eventID="1011"] BOMAn application event log entry... """ PRI -- 165 VERSION -- 1 TIMESTAMP -- 2019-07-11T22:14:15.003Z HOSTNAME -- aliyun.example.com APP-NAME -- evntslog PROCID -- "-" MSGID -- ID47 STRUCTURED-DATA -- [exampleSDID@32473 iut="3" eventSource="Application" eventID="1011"] MESSAGE -- An application event log entry... -
RFC 3164 protocol
The RFC 3164 protocol includes the following fields. For more information, see the official protocol document.
PRI HEADER[TIME HOSTNAME] MSGThe following example describes these fields:
""" <30>Oct 9 22:33:20 hlfedora auditd[1787]: The audit daemon is exiting. """ PRI -- 30 HEADER - TIME -- Oct 9 22:33:20 - HOSTNAME -- hlfedora MSG - TAG -- auditd[1787] - Content -- The audit daemon is exiting.
Use GROK to parse common Syslog formats
You can use GROK to parse several common Syslog formats. For more information about GROK rules, see GROK pattern reference.
-
Parse the TraditionalFormat format
-
Raw log
receive_time: 1558663265 __topic__: content: May 5 10:20:57 iZbp1a65x3r1vhpe94fi2qZ systemd: Started System Logging Service. -
SLS DSL rule
e_regex('content', grok('%{SYSLOGBASE} %{GREEDYDATA:message}')) -
Processing result
receive_time: 1558663265 __topic__: content: May 5 10:20:57 iZbp1a65x3r1vhpe94fi2qZ systemd: Started System Logging Service. timestamp: May 5 10:20:57 logsource: iZbp1a65x3r1vhpe94fi2qZ program: systemd message: Started System Logging Service.
-
-
Parse the FileFormat
-
Raw log
receive_time: 1558663265 __topic__: content: 2019-05-06T09:26:07.874593+08:00 iZbp1a65x3r1vhpe94fi2qZ user: 834753 -
SLS DSL rule
e_regex('content',grok('%{TIMESTAMP_ISO8601:timestamp} %{SYSLOGHOST:hostname} %{SYSLOGPROG} %{GREEDYDATA:message}')) -
Processing result
receive_time: 1558663265 __topic__: content: 2019-05-06T09:26:07.874593+08:00 iZbp1a65x3r1vhpe94fi2qZ user: 834753 timestamp: 2019-05-06T09:26:07.874593+08:00 hostname: iZbp1a65x3r1vhpe94fi2qZ program: user message: 834753
-
-
Parse logs in RSYSLOG_SyslogProtocol23Format
-
Raw log
receive_time: 1558663265 __topic__: content: <13>1 2019-05-06T11:50:16.015554+08:00 iZbp1a65x3r1vhpe94fi2qZ user - - - twish -
SLS DSL rule
e_regex('content',grok('%{POSINT:priority}>%{NUMBER:version} %{TIMESTAMP_ISO8601:timestamp} %{SYSLOGHOST:hostname} %{PROG:program} - - - %{GREEDYDATA:message}')) -
Processing result
receive_time: 1558663265 __topic__: content: <13>1 2019-05-06T11:50:16.015554+08:00 iZbp1a65x3r1vhpe94fi2qZ user - - - twish priority: 13 version: 1 timestamp: 2019-05-06T11:50:16.015554+08:00 hostname: iZbp1a65x3r1vhpe94fi2qZ program: user message: twish
-
-
Parse the RSYSLOG_DebugFormat format
-
Raw log
receive_time: 1558663265 __topic__: content: 2019-05-06T14:29:37.558854+08:00 iZbp1a65x3r1vhpe94fi2qZ user: environment -
SLS DSL rule
e_regex('content',grok('%{TIMESTAMP_ISO8601:timestamp} %{SYSLOGHOST:hostname} %{SYSLOGPROG} %{GREEDYDATA:message}')) -
Processing result
receive_time: 1558663265 __topic__: content: 2019-05-06T14:29:37.558854+08:00 iZbp1a65x3r1vhpe94fi2qZ user: environment timestamp: 2019-05-06T14:29:37.558854+08:00 hostname: iZbp1a65x3r1vhpe94fi2qZ program: user message: environment
-
Use GROK to parse uncommon Syslog log formats
You can use GROK to parse two uncommon Syslog formats used by Fluentd: FluentRFC5424 and FluentRFC3164.
-
FluentRFC5424 format
-
Raw log
receive_time: 1558663265 __topic__: content: <16>1 2019-02-28T12:00:00.003Z 192.168.0.1 aliyun 11111 ID24224 [exampleSDID@20224 iut='3' eventSource='Application' eventID='11211] Hi, from Fluentd! -
SLS DSL rule
e_regex('content',grok('%{POSINT:priority}>%{NUMBER:version} %{TIMESTAMP_ISO8601:timestamp} %{SYSLOGHOST:hostname} %{WORD:ident} %{USER:pid} %{USERNAME:msgid} (?P<extradata>(\[(.*)\]|[^ ])) %{GREEDYDATA:message}')) -
Processing result
receive_time: 1558663265 __topic__: content: <16>1 2019-02-28T12:00:00.003Z 192.168.0.1 aliyun 11111 ID24224 [exampleSDID@20224 iut='3' eventSource='Application' eventID='11211] Hi, from aliyun! priority: 16 version: 1 timestamp: 2019-02-28T12:00:00.003Z hostname: 192.168.0.1 ident: aliyun pid: 11111 msgid: ID24224 extradata: [exampleSDID@20224 iut='3' eventSource='Application' eventID='11211] message: Hi, from aliyun!
-
-
FluentRFC3164 format
-
Raw log
receive_time: 1558663265 __topic__: content: <6>Feb 28 12:00:00 192.168.0.1 aliyun[11111]: [error] Syslog test -
SLS DSL rule
e_regex('content', grok('%{POSINT:priority}>%{SYSLOGTIMESTAMP:timestamp} %{SYSLOGHOST:hostname} %{WORD:ident}(?P<pid>(\[[a-zA-Z0-9._-]+\]|[^:])): (?P<level>(\[(\w+)\]|[^ ])) %{GREEDYDATA:message}')) -
Processing result
receive_time: 1558663265 __topic__: content: <6>Feb 28 12:00:00 192.168.0.1 aliyun[11111]: [error] Syslog test priority: 6 timestamp: Feb 28 12:00:00 hostname: 192.168.0.1 ident: aliyun pid: [11111] level: [error] message: Syslog test
-
-
Extended priority parsing
After you parse Syslog logs in FluentRFC5424 or FluentRFC3164 format, you can further parse the priority field to extract facility and severity information. For more information, see e_syslogrfc. The following is an example:
-
Raw log
receive_time: 1558663265 __topic__: content: <13>1 2019-05-06T11:50:16.015554+08:00 iZbp1a65x3r1vhpe94fi2qZ user - - - twish priority: 13 version: 1 timestamp: 2019-05-06T11:50:16.015554+08:00 hostname: iZbp1a65x3r1vhpe94fi2qZ program: user message: twish -
SLS DSL rule
e_syslogrfc("priority","SYSLOGRFC5424") -
Processing result
receive_time: 1558663265 __topic__: content: <13>1 2019-05-06T11:50:16.015554+08:00 iZbp1a65x3r1vhpe94fi2qZ user - - - twish priority: 13 version: 1 timestamp: 2019-05-06T11:50:16.015554+08:00 hostname: iZbp1a65x3r1vhpe94fi2qZ program: user message: twish _facility_: 1 _severity_: 5 _severitylabel_: Notice: normal but significant condition _facilitylabel_: user-level messages
-