摘要: Logstash配置总结和实例
这里记录Logstash配置中注意的事项：
整个配置文件分为三部分：input,filter,output。参考这里的介绍 https://www.elastic.co/guide/en/logstash/current/configuration-file-structure.html
1 在Windows中，文件路径中分隔符要使用/而不是\。如果使用了\，那么*匹配将会失败。
2 默认的@timestamp是使用UTC时间表示的，所以对于北京时间会有8小时差，我们很多情况会使用日期来做type区分日志，这时有个办法是在filter中增加ruby来做转换。其实是新加了一个属性来表示日期。
另外有个选项是在date中有timezone的属性，但是如果设置它为UTC，那么整个日志中的时间就是UTC时间了，不是一个理想的做法。
下面是一个配置解释
2016-5-5更新：
#整个配置文件分为三部分：input,filter,output
#参考这里的介绍 https://www.elastic.co/guide/en/logstash/current/configuration-file-structure.html
input {
#file可以多次使用，也可以只写一个file而设置它的path属性配置多个文件实现多文件监控
file {
#type是给结果增加了一个属性叫type值为"<xxx>"的条目。这里的type，对应了ES中index中的type，即如果输入ES时，没有指定type，那么这里的type将作为ES中index的type。
type => "apache-access"
path => "/apphome/ptc/Windchill_10.0/Apache/logs/access_log*"
#start_position可以设置为beginning或者end，beginning表示从头开始读取文件，end表示读取最新的，这个也要和ignore_older一起使用。
start_position => beginning
#sincedb_path表示文件读取进度的记录，每行表示一个文件，每行有两个数字，第一个表示文件的inode，第二个表示文件读取到的位置（byteoffset）。默认为$HOME/.sincedb*
sincedb_path => "/opt/logstash-2.3.1/sincedb_path/access_progress"
#ignore_older表示了针对多久的文件进行监控，默认一天，单位为秒，可以自己定制，比如默认只读取一天内被修改的文件。
ignore_older => 604800
#add_field增加属性。这里使用了${HOSTNAME}，即本机的环境变量，如果要使用本机的环境变量，那么需要在启动命令上加--alow-env。
add_field => {"log_hostname"=>"${HOSTNAME}"}
#这个值默认是\n 换行符，如果设置为空""，那么后果是每个字符代表一个event
delimiter => ""
#这个表示关闭超过（默认）3600秒后追踪文件。这个对于multiline来说特别有用。... 这个参数和logstash对文件的读取方式有关，两种方式read tail，如果是read
close_older => 3600
coodec => multiline {
pattern => "^\s"
#这个negate是否定的意思，意思跟pattern相反，也就是不满足patter的意思。
# negate => ""
#what有两个值可选 previous和next，举例说明，java的异常从第二行以空格开始，这里就可以pattern匹配空格开始，what设置为previous意思是空格开头这行跟上一行属于同一event。另一个例子，有时候一条命令太长，当以\结尾时表示这行属于跟下一行属于同一event，这时需要使用negate=>true，what=>'next'。
what => "previous"
auto_flush_interval => 60
}
}
file {
type => "methodserver-log"
path => "/apphome/ptc/Windchill_10.0/Windchill/logs/MethodServer-1604221021-32380.log"
start_position => beginning
sincedb_path => "/opt/logstash-2.3.1/sincedb_path/methodserver_process"
# ignore_older => 604800
}
}
filter{
#执行ruby程序，下面例子是将日期转化为字符串赋予daytag
ruby {
code => "event['daytag'] = event.timestamp.time.localtime.strftime('%Y-%m-%d')"
}
# if [path] =~ "access" {} else if [path] =~ "methodserver" {} else if [path] =~ "servermanager" {} else {} 注意语句结构
if [path] =~ "MethodServer" { #z这里的=~是匹配正则表达式
grok {
patterns_dir => ["/opt/logstash-2.3.1/patterns"] #自定义正则匹配
# Tue 4/12/16 14:24:17: TP-Processor2: hirecode---->77LS
match => { "message" => "%{DAY:log_weekday} %{DATE_US:log_date} %{TIME:log_time}: %{GREEDYDATA:log_data}"}
}
#mutage是做转换用的
mutate {
replace => { "type" => "apache" } #替换属性值
convert => { #类型转换
"bytes" => "integer" #例如还有float
"duration" => "integer"
"state" => "integer"
}
#date主要是用来处理文件内容中的日期的。内容中读取的是字符串，通过date将它转换为@timestamp。参考https://www.elastic.co/guide/en/logstash/current/plugins-filters-date.html#plugins-filters-date-match
# date {
# match => [ "logTime" , "dd/MMM/yyyy:HH:mm:ss Z" ]
# }
}else if [type] in ['tbg_qas','mbg_pre'] { # if ... else if ... else if ... else结构
}else {
drop{} # 将event丢弃
}
}
output {
stdout{ codec=>rubydebug} # 直接输出，调试用起来方便
# 输出到redis
redis {
host => '10.120.20.208'
data_type => 'list'
key => '10.99.201.34:access_log_2016-04'
}
# 输出到ES
elasticsearch {
hosts =>"192.168.0.15:9200"
index => "%{sysid}_%{type}"
document_type => "%{daytag}"
}
}下面是两个真实的例子，第一个是从应用到redis，第二个是从redis到ES。
input {
file {
type => "log_raw_data"
path => "/apphome/ptc/Windchill_10.0/Windchill/logs/gc/*GC.log"
start_position => end
sincedb_path => "/opt/logstash-2.3.1/sincedb_path/log_progress"
# ignore_older => 604800
add_field => {"sysid"=>"tbg_qas"}
}
file {
type => "log_raw_data"
path => ["/apphome/ptc/Windchill_10.0/Windchill/logs/*MethodServer*.log","/apphome/ptc/Windchill_10.0/Windchill/logs/ServerManager-*.log"]
start_position => end
sincedb_path => "/opt/logstash-2.3.1/sincedb_path/log_progress"
# ignore_older => 604800
add_field => {"sysid"=>"tbg_qas"}
close_older => 60
codec => multiline {
# patterns_dir => ["D:/app/logstash-2.3.1/patterns"]
pattern => "^%{DAY} %{DATESTAMP}:"
negate => true
what => "previous"
# auto_flush_interval => 20
}
}
}
output {
# stdout{ codec=>rubydebug}
redis {
host => '10.120.20.208'
data_type => 'list'
key => 'log_raw_data'
}
redis {
host => '10.120.31.142'
data_type => 'list'
key => 'log_raw_data'
}
}input {
redis {
host => "localhost"
data_type => "list"
port => "6379"
key => "log_raw_data"
type => "redis-input"
}
}
filter{
ruby {
code => "event['daytag'] = event.timestamp.time.localtime.strftime('%Y-%m-%d')"
}
if [path] =~ "access" {
grok {
match => { "message" => "%{IPORHOST:clientip} %{HTTPDUSER:ident} %{USER:username} \[%{HTTPDATE:logtime}\] \"%{WORD:verb} %{NOTSPACE:request} (?:%{NOTSPACE:httpversion}|)\" (?:%{NUMBER:state}|-) (?:%{NUMBER:bytes}|-) %{NUMBER:duration}"}
}
mutate {
replace => { "type" => "apache" }
convert => {
"bytes" => "integer"
"duration" => "integer"
"state" => "integer"
}
}
date {
match => [ "logtime" , "dd/MMM/yyyy:HH:mm:ss Z" ]
}
}else if [path] =~ ".*ServerManager.*GC\.log" {
if [message] =~ "\[Full GC" {
grok {
match => {"message" => "%{TIMESTAMP_ISO8601:logtime}: %{GREEDYDATA:gcdetail} \[Times: user=%{BASE10NUM:usertime} sys=%{BASE10NUM:systime}, real=%{BASE10NUM:realtime} secs\]"}
}
date {
match => ["logtime" , "yyyy-MM-dd'T'HH:mm:ss.SSS'+0800'"]
}
}else if [message] =~ "\[GC" {
grok {
match => {"message" => "%{TIMESTAMP_ISO8601:logtime}: %{GREEDYDATA:gcdetail} \[Times: user=%{BASE10NUM:usertime} sys=%{BASE10NUM:systime}, real=%{BASE10NUM:realtime} secs\]"}
}
date {
match => ["logtime" , "yyyy-MM-dd'T'HH:mm:ss.SSS'+0800'"]
}
}else{
drop {}
}
mutate {
replace => {"type" => "smgc" }
convert => {
"usertime" => "float"
"systime" => "float"
"realtime" => "float"
}
}
}else if [path] =~ ".*MethodServer.*GC\.log" {
if [message] =~ "\[Full GC" {
grok {
match => {"message" => "%{TIMESTAMP_ISO8601:logtime}: %{GREEDYDATA:gcdetail} \[Times: user=%{BASE10NUM:usertime} sys=%{BASE10NUM:systime}, real=%{BASE10NUM:realtime} secs\]"}
}
date {
match => ["logtime" , "yyyy-MM-dd'T'HH:mm:ss.SSS'+0800'"]
}
}else if [message] =~ "\[GC" {
grok {
match => {"message" => "%{TIMESTAMP_ISO8601:logtime}: %{GREEDYDATA:gcdetail} \[Times: user=%{BASE10NUM:usertime} sys=%{BASE10NUM:systime}, real=%{BASE10NUM:realtime} secs\]"}
}
date {
match => ["logtime" , "yyyy-MM-dd'T'HH:mm:ss.SSS'+0800'"]
}
}else{
drop {}
}
mutate {
replace => {"type" => "msgc" }
convert => {
"usertime" => "float"
"systime" => "float"
"realtime" => "float"
}
}
}else if [path] =~ "MethodServer" {
grok {
match => { "message" => "%{DAY:weekday} %{DATESTAMP:logtime}: %{GREEDYDATA:logdata}"}
}
date {
match => [ "logtime" , "M/d/yy HH:mm:ss" ]
}
mutate { replace => { "type" => "ms" } }
}else if [path] =~ "ServerManager" {
grok {
match => { "message" => "%{DAY:weekday} %{DATESTAMP:logtime}: %{GREEDYDATA:logdata}"}
}
date {
match => [ "logtime" , "M/d/yy HH:mm:ss" ]
}
mutate { replace => { "type" => "sm" } }
}else if [path] =~ "Process_Archive" {
grok {
patterns_dir => ["/opt/logstash-2.3.1/patterns"]
match => { "message" => "%{PROCESS_DATETIME:logtime} %{GREEDYDATA:logdata}"}
}
date {
match => [ "logtime" , "yyyy MMM dd HH:mm:ss:SSS 'GMT +8'" ]
}
mutate { replace => { "type" => "prc_arc" } }
}else if [path] =~ "ESISAPAdapterConfiguration" {
grok {
patterns_dir => ["/opt/logstash-2.3.1/patterns"]
match => { "message" => "%{PROCESS_DATETIME:logtime} %{GREEDYDATA:logdata}"}
}
date {
match => [ "logtime" , "yyyy MMM dd HH:mm:ss:SSS 'GMT +8'" ]
}
mutate { replace => { "type" => "esi_adp" } }
}else if [path] =~ "LenovoAdapterConfiguration" {
grok {
patterns_dir => ["/opt/logstash-2.3.1/patterns"]
match => { "message" => "%{PROCESS_DATETIME:logtime} %{GREEDYDATA:logdata}"}
}
date {
match => [ "logtime" , "yyyy MMM dd HH:mm:ss:SSS 'GMT +8'" ]
}
mutate { replace => { "type" => "le_adp" } }
}else {
mutate { replace => { "type" => "other" } }
# drop {}
}
# extractnumbers {
# source => "duration"
# }
}
output {
# stdout{ codec=>rubydebug}
elasticsearch {
hosts =>"192.168.0.15:9200"
index => "%{sysid}_%{type}"
document_type => "%{daytag}"
}
}
© 著作权归作者所有
扫一扫
279
被折叠的 条评论
为什么被折叠?
到【灌水乐园】发言
