现在大部分公司都会选择将应用、中间件、系统等日志存储在 Elasticsearch 中，如何发现日志中的异常数据并且及时告警通知就显得十分重要。本文将会介绍两种主流的日志监控方案，分别是 Yelp 公司开源的 ElastAlert 和 Elastic 官方的商业版功能 Watcher。
如下图所示，日志数据源是一台 Nginx 服务器，在该服务器上安装 Filebeat 收集 Nginx 日志并输出到 Elasticsearch，之后会分别演示用 ElastAlert 和 Watcher 两种方案监控日志并进行告警。
部署 Nginx
安装依赖
代码语言：javascript复制yum install -y gcc gcc-c autoconf pcre pcre-devel make automake wget httpd-tools vim tree zlib-devel
下载安装包
代码语言：javascript复制wget http://nginx.org/download/nginx-1.14.0.tar.gz
tar -xzvf nginx-1.14.0.tar.gz
编译安装
代码语言：javascript复制cd nginx-1.14.0
./configure
配置 Nginx
编辑配置文件 /usr/local/nginx/conf/nginx.conf，在 Nginx 上配置一个静态网页服务。
代码语言：javascript复制worker_processes 1;
events {
worker_connections 1024;
}
http {
server {
listen 80;
location / {
root html;
}
}
}
启动 Nginx：
代码语言：javascript复制sbin/nginx
访问 Nginx：
部署 Filebeat
下载并安装 Filebeat。
代码语言：javascript复制curl -L -O https://artifacts.elastic.co/downloads/beats/filebeat/filebeat-7.14.0-x86_64.rpm
sudo rpm -vi filebeat-7.14.0-x86_64.rpm
编辑 /etc/filebeat/filebeat.yml 配置文件，读取 Nginx 日志文件输出到 Elasticsearch 的 nginx 索引中，后缀是当前日期。
代码语言：javascript复制filebeat.inputs:
- type: log
enabled: true
paths:
- /usr/local/nginx/logs/*.log
output.elasticsearch:
hosts: ["192.168.1.8:9200"]
index: "nginx-%{ yyyy.MM.dd}"
#username: "elastic"
#password: "changeme"
setup.ilm.enabled: false
setup.template.name: "nginx"
setup.template.pattern: "nginx-*"
启动 Filebeat：
代码语言：javascript复制systemctl start filebeat
ElastAlert
ElastAlert 是 Yelp 公司开源的一套用 Python 写的 Elasticsearch 告警框架，可以从 Elasticsearch 当中查询出匹配规则的数据进行告警。
ElastAlert 有以下特点：
- 支持多种匹配规则（频率、阈值、数据变化、黑白名单、变化率等）。
- 支持多种告警类型（邮件、HTTP POST、自定义脚本等）。
- 支持用户自定义规则和告警类型。
- 匹配项汇总报警，重复告警抑制，告警失败重试和过期。
- 可用性强，状态信息保存到 Elasticsearch 的索引中。
- 支持调试和审计。
部署 Elastalert
安装 Python
代码语言：javascript复制wget https://www.python.org/ftp/python/3.6.9/Python-3.6.9.tgz
tar -zxvf Python-3.6.9.tgz
cd Python-3.6.9
./configure
make && make install
检查 Python 版本：
代码语言：javascript复制python3 -V
安装依赖
代码语言：javascript复制yum install gcc libffi-devel python3-devel openssl-devel -y
pip3 install -U pip
pip3 install "setuptools>=11.3"
安装 Elastalert
代码语言：javascript复制python3 install elastalert
配置 Elastalert
克隆代码到本地：
代码语言：javascript复制git clone https://github.com/Yelp/elastalert.git
cd elastalert
我们可以在 ElastAlert 源码文件的根目录下找到一个叫做 config.yaml.example 的文件，修改文件名为 config.yaml：
代码语言：javascript复制mv config.yaml.example config.yaml
创建存放规则的目录。
代码语言：javascript复制mkdir rules
cd rules
编辑 config.yaml 文件，修改主配置：
代码语言：javascript复制#规则存放的目录
rules_folder: rules
#运行的频率
run_every:
minutes: 1
#ElastAlert 将缓存最近一段时间的结果，以防某些日志源不是实时的
buffer_time:
minutes: 45
#Elasticsearch 地址
es_host: 192.168.1.8
#Elasticsearch 端口
es_port: 9200
#Elasticsearch 用户名密码（可选）
#es_username: someusername
#es_password: somepassword
#ElastAlert 元数据存储索引
writeback_index: elastalert_status
#如果警报因某种原因失败，ElastAlert将重试发送警报，直到该时间段结束
alert_time_limit:
days: 2
创建 rules/nginx.yaml 文件，编辑 rule：
规则内容为：在 1 分钟内如果查询 nginx-* 索引的 message 字段匹配 到 error 5 次就触发告警，往指定的 URL 发送一个 HTTP POST 请求。
代码语言：javascript复制# Alert when the rate of events exceeds a threshold
# (Required)
# Elasticsearch host
es_host: 192.168.1.8
# (Required)
# Elasticsearch port
es_port: 9200
# (OptionaL) Connect with SSL to elasticsearch
#use_ssl: True
# (Optional) basic-auth username and password for elasticsearch
#es_username: someusername
#es_password: somepassword
# (Required)
# Rule name, must be unique
name: nginx rule
# (Required)
# Type of alert.
# the frequency rule type alerts when num_events events occur with timeframe time
type: frequency
# (Required)
# Index to search, wildcard supported
index: nginx-*
# (Required, frequency specific)
# Alert when this many documents matching the query occur within a timeframe
num_events: 5
# (Required, frequency specific)
# num_events must occur within this amount of time to trigger an alert
timeframe:
minutes: 1
# (Required)
# A list of elasticsearch filters used for find events
# These filters are joined with AND and nested in a filtered query
# For more info: http://www.elasticsearch.org/guide/en/elasticsearch/reference/current/query-dsl.html
filter:
- term:
message: "error"
# (Required)
# The alert is use when a match is found
alert:
- "post"
http_post_url: "https://webhook.site/2f64f4b3-8b43-488c-b2df-695136079e36"
https://webhook.site 网站提供了测试的 Webhook 接口，每个人的 URL 都是独立的，拷贝这个 URL 复制到 http_post_url 中。
ElastAlert 会把执行记录存放到一个索引中，可以方便我们审核和调试。使用以下命令创建这个索引的，默认情况下，索引名叫 elastalert_status。
代码语言：javascript复制root@ydt-net-es-node1:/software #elastalert-create-index
Enter Elasticsearch host: 192.168.1.8
Enter Elasticsearch port: 9200
Use SSL? t/f: f
#如果有认证输入用户名密码
Enter optional basic-auth username (or leave blank):
Enter optional basic-auth password (or leave blank):
Enter optional Elasticsearch URL prefix (prepends a string to the URL of every request):
New index name? (Default elastalert_status)
New alias name? (Default elastalert_alerts)
Name of existing index to copy? (Default None)
Elastic Version: 7.9.3
Reading Elastic 6 index mappings:
Reading index mapping 'es_mappings/6/silence.json'
Reading index mapping 'es_mappings/6/elastalert_status.json'
Reading index mapping 'es_mappings/6/elastalert.json'
Reading index mapping 'es_mappings/6/past_elastalert.json'
Reading index mapping 'es_mappings/6/elastalert_error.json'
New index elastalert_status created
Done!
发送 2 个请求，1 个是正确请求，1 个是错误请求。
代码语言：javascript复制> curl http://192.168.1.134 -I
HTTP/1.1 200 OK
Server: nginx/1.14.2
Date: Mon, 16 Aug 2021 07:28:42 GMT
Content-Type: text/html
Content-Length: 612
Last-Modified: Wed, 16 Jun 2021 02:46:13 GMT
Connection: keep-alive
ETag: "60c965f5-264"
Accept-Ranges: bytes
> curl http://192.168.1.134/xxxxxx -I
HTTP/1.1 404 Not Found
Server: nginx/1.14.2
Date: Mon, 16 Aug 2021 07:28:43 GMT
Content-Type: text/html
Content-Length: 169
Connection: keep-alive
在 Kibana 上可以看到 Nginx 的日志，错误请求会在 access.log 和 error.log 各写一次，因此这里看到 3 条记录。
运行 elastalert-test-rule 命令检验配置文件是否正确并且可以看到规则匹配的次数，elastalert-test-rule 命令并不会真正触发告警。
代码语言：javascript复制> elastalert-test-rule rules/nginx.yaml
INFO:elastalert:Note: In debug mode, alerts will be logged to console but NOT actually sent.
To send them but remain verbose, use --verbose instead.
Didn't get any results.
INFO:elastalert:Note: In debug mode, alerts will be logged to console but NOT actually sent.
To send them but remain verbose, use --verbose instead.
1 rules loaded
INFO:apscheduler.scheduler:Adding job tentatively -- it will be properly scheduled when the scheduler starts
#匹配一次
INFO:elastalert:Queried rule nginx rule from 2021-08-16 15:28 CST to 2021-08-16 15:29 CST: 1 / 1 hits
Would have written the following documents to writeback index (default is elastalert_status):
elastalert_status - {'rule_name': 'nginx rule', 'endtime': datetime.datetime(2021, 8, 16, 7, 29, 30, 422431, tzinfo=tzutc()), 'starttime': datetime.datetime(2021, 8, 16, 7, 28, 29, 822431, tzinfo=tzutc()), 'matches': 0, 'hits': 1, '@timestamp': datetime.datetime(2021, 8, 16, 7, 29, 30, 527080, tzinfo=tzutc()), 'time_taken': 0.02203655242919922}
1分钟内连续发送错误请求 5 次达到触发告警的阈值：
代码语言：javascript复制for i in {1..3};do curl http://192.168.1.134/xxxxxx -I;done
此时可以看到发送的告警格式。
代码语言：javascript复制> elastalert-test-rule rules/nginx.yaml
INFO:elastalert:Note: In debug mode, alerts will be logged to console but NOT actually sent.
To send them but remain verbose, use --verbose instead.
Didn't get any results.
INFO:elastalert:Note: In debug mode, alerts will be logged to console but NOT actually sent.
To send them but remain verbose, use --verbose instead.
1 rules loaded
INFO:apscheduler.scheduler:Adding job tentatively -- it will be properly scheduled when the scheduler starts
INFO:elastalert:Queried rule nginx rule from 2021-08-16 15:33 CST to 2021-08-16 15:34 CST: 5 / 5 hits
INFO:elastalert:Alert for nginx rule at 2021-08-16T07:34:26.230Z:
INFO:elastalert:nginx rule
At least 5 events occurred between 2021-08-16 15:33 CST and 2021-08-16 15:34 CST
@timestamp: 2021-08-16T07:34:26.230Z
_id: 0CDiTXsBCANUjLffFM2O
_index: nginx-2021.08.16
_type: _doc
agent: {
"ephemeral_id": "4ee4bd89-cb8e-43fb-9331-476c229a5480",
"hostname": "nginx-plus1",
"id": "629442a8-34ab-40db-80a8-16e4fda8dec7",
"name": "nginx-plus1",
"type": "filebeat",
"version": "7.14.0"
}
ecs: {
"version": "1.10.0"
}
host: {
"name": "nginx-plus1"
}
input: {
"type": "log"
}
log: {
"file": {
"path": "/usr/local/nginx/logs/error.log"
},
"offset": 16944
}
message: 2021/08/16 15:34:22 [error] 4022#0: *40 open() "/usr/local/nginx/html/xxxxxx" failed (2: No such file or directory), client: 192.168.1.35, server: , request: "GET /xxxxxx HTTP/1.1", host: "192.168.1.134"
num_hits: 5
num_matches: 1
Would have written the following documents to writeback index (default is elastalert_status):
silence - {'exponent': 0, 'rule_name': 'nginx rule', '@timestamp': datetime.datetime(2021, 8, 16, 7, 34, 42, 866184, tzinfo=tzutc()), 'until': datetime.datetime(2021, 8, 16, 7, 35, 42, 866174, tzinfo=tzutc())}
elastalert_status - {'rule_name': 'nginx rule', 'endtime': datetime.datetime(2021, 8, 16, 7, 34, 42, 810992, tzinfo=tzutc()), 'starttime': datetime.datetime(2021, 8, 16, 7, 33, 42, 210992, tzinfo=tzutc()), 'matches': 1, 'hits': 5, '@timestamp': datetime.datetime(2021, 8, 16, 7, 34, 42, 868045, tzinfo=tzutc()), 'time_taken': 0.015259981155395508}
使用以下命令运行 elastalert，可以看到触发了告警：
代码语言：javascript复制> elastalert --verbose --rule rules/nginx.yaml
1 rules loaded
INFO:elastalert:Starting up
INFO:elastalert:Disabled rules are: []
INFO:elastalert:Sleeping for 59.999839 seconds
INFO:elastalert:Queried rule nginx rule from 2021-08-16 14:54 CST to 2021-08-16 15:39 CST: 7 / 7 hits
INFO:elastalert:HTTP Post alert sent.
INFO:elastalert:Ran nginx rule from 2021-08-16 14:54 CST to 2021-08-16 15:39 CST: 7 query hits (0 already seen), 1 matches, 1 alerts sent
访问 https://webhook.site 网站可以看到 ElastAlert 发送的 HTTP POST 请求。
查询 elastalert_status 索引可以看到 ElastAlert 的执行记录。
代码语言：javascript复制GET elastalert_status/_search
#返回结果
{
"took" : 0,
"timed_out" : false,
"_shards" : {
"total" : 1,
"successful" : 1,
"skipped" : 0,
"failed" : 0
},
"hits" : {
"total" : {
"value" : 1,
"relation" : "eq"
},
"max_score" : 1.0,
"hits" : [
{
"_index" : "elastalert_status",
"_type" : "_doc",
"_id" : "1SDmTXsBCANUjLff0M1Q",
"_score" : 1.0,
"_source" : {
"match_body" : {
"input" : {
"type" : "log"
},
"agent" : {
"hostname" : "nginx-plus1",
"name" : "nginx-plus1",
"id" : "629442a8-34ab-40db-80a8-16e4fda8dec7",
"ephemeral_id" : "4ee4bd89-cb8e-43fb-9331-476c229a5480",
"type" : "filebeat",
"version" : "7.14.0"
},
"@timestamp" : "2021-08-16T07:34:26.230Z",
"ecs" : {
"version" : "1.10.0"
},
"log" : {
"file" : {
"path" : "/usr/local/nginx/logs/error.log"
},
"offset" : 16740
},
"host" : {
"name" : "nginx-plus1"
},
"message" : "2021/08/16 15:34:22 [error] 4022#0: *39 open() "/usr/local/nginx/html/xxxxxx" failed (2: No such file or directory), client: 192.168.1.35, server: , request: "GET /xxxxxx HTTP/1.1", host: "192.168.1.134"",
"_id" : "zyDiTXsBCANUjLffFM2O",
"_index" : "nginx-2021.08.16",
"_type" : "_doc",
"num_hits" : 7,
"num_matches" : 1
},
"rule_name" : "nginx rule",
"alert_info" : {
"type" : "http_post",
"http_post_webhook_url" : [
"https://webhook.site/2f64f4b3-8b43-488c-b2df-695136079e36"
]
},
"alert_sent" : true,
"alert_time" : "2021-08-16T07:39:35.185929Z",
"match_time" : "2021-08-16T07:34:26.230Z",
"@timestamp" : "2021-08-16T07:39:37.418536Z"
}
}
]
}
}
Watcher
Watcher 是 Elastic 官方提供的一个对日志数据监控和报警的功能，Watcher 属于收费功能，我们可以在 License Management 中开启 30 天的试用。
Watcher 由以下 5 个部分组成：
- trigger：定义 watcher 触发的时间或者周期。
- input：定义数据的来源，可以是一个索引或者 HTTP 请求的结果等等。如果没有设置输入将为空。
- condition：定义执行 action 触发的条件。如果没有设置默认总是触发 action。
- transform（可选）：修改 watcher 的 payload。
- actions：定义执行的动作，例如 email,webhook,index,logging,slack 等等。
创建 1 个 Watcher：
- trigger：每分钟运行一次。
- input：通配符匹配 nginx-* 的索引，查询 message 字段中的 error 关键字，每次针对在过去5分钟内发生的事件来进行查询。
- condition：如果在查询结果中，匹配到 1 次，就触发 action。
- action：向指定 URL 发送一个 HTTP POST 请求。
PUT _watcher/watch/nginx-watcher
{
"trigger": {
"schedule" : {
"interval" : "1m"
}
},
"input": {
"search": {
"request": {
"indices": [
"nginx-*"
],
"body": {
"query": {
"bool": {
"must": {
"match": {
"message": "error"
}
},
"filter": {
"range": {
"@timestamp": {
"from": "{{ctx.trigger.scheduled_time}}||-5m",
"to": "{{ctx.trigger.triggered_time}}"
}
}
}
}
}
}
}
}
},
"condition": {
"compare": {
"ctx.payload.hits.total": {
"gt": 0
}
}
},
"actions": {
"my_webhook": {
"throttle_period": "2m",
"webhook": {
"method": "POST",
"url": "https://webhook.site/2f64f4b3-8b43-488c-b2df-695136079e36",
"body": "Number of Nginx Error: {{ctx.payload.hits.total}}"
}
}
}
}
查看刚刚创建的 watcher：
1分钟内连续发送 5 次错误请求。
代码语言：javascript复制for i in {1..3};do curl http://192.168.1.134/xxxxxx -I;done
查看 watcher 状态，可以看到触发了 action。
访问 https://webhook.site 可以看到最新的 Webhook 事件已经被触发了，而且它的 Raw Content 和我们之前定义的 body 格式是一致的。
如果我们设置的 watcher 间隔时间比较久，Elasticsearch 为了方便我们测试，提供了_execute 接口，通过执行下面命令可以立即运行一下我们的 watcher。
代码语言：javascript复制PUT _watcher/watch/nginx-watcher/_execute
参考资料
- https://zhuanlan.zhihu.com/p/386722918
- https://elastalert.readthedocs.io/
- https://www.elastic.co/guide/en/elasticsearch/reference/7.14/xpack-alerting.html
- https://blog.csdn.net/UbuntuTouch/article/details/106298651
- https://elasticstack.blog.csdn.net/article/details/105340379
- https://elasticstack.blog.csdn.net/article/details/103820572
