一、安装elastalert
1、环境
- CentOS:7.4
- Python:3.6.9
- pip:19.3
- elastalert:0.2.1
- elk:7.3.2
2、配置Python3.6.9环境
安装依赖包
yum -y install wget openssl openssl-devel gcc gcc-c++
下载包
wget https://www.python.org/ftp/python/3.6.9/Python-3.6.9.tgz
安装
mkdir -p /usr/local/python
tar xf Python-3.6.9.tgz
cd Python-3.6.9
./configure --prefix=/usr/local/python --with-openssl
make && make install
配置
mv /usr/bin/python /usr/bin/python_old
ln -s /usr/local/python/bin/python3 /usr/bin/python
ln -s /usr/local/python/bin/pip3 /usr/bin/pip
pip install --upgrade pip
注意,所有依赖python2的脚本或者命令,需要更改为python2.7,因为现在默认的python版本为3.6,例如
sed -i '1s/python/python2.7/g' /usr/bin/yum
sed -i '1s/python/python2.7/g' /usr/libexec/urlgrabber-ext-down
验证
$ python -V
Python 3.6.9
$ pip -V
pip 19.3 from /usr/local/python/lib/python3.6/site-packages/pip (python 3.6)
3、安装elastalert
下载包
git clone https://github.com/Yelp/elastalert.git
cd elastalert
安装
pip install "elasticsearch<7,>6"
pip install -r requirements.txt
python setup.py install
安装成功后可以看到四个命令
ll /usr/local/python/bin/elastalert*
/usr/local/python/bin/elastalert
/usr/local/python/bin/elastalert-create-index
/usr/local/python/bin/elastalert-rule-from-kibana
/usr/local/python/bin/elastalert-test-rule
软连接到/usr/bin下,方便使用
ln -s /usr/local/python/bin/elastalert* /usr/bin
- elastalert 报警执行的命令,会根据报警规则执行相应操作。
- elastalert-create-index会创建一个索引,ElastAlert会把执行记录存放到这个索引中,默认情况下,索引名叫elastalert_status。其中有4个_type,都有自己的@timestamp字段,所以同样也可以用kibana来查看这个索引的日志记录情况。
- elastalert-rule-from-kibana从Kibana3已保存的仪表盘中读取Filtering设置,帮助生成config.yaml里的配置。不过注意,它只会读取filtering,不包括queries。
- elastalert-test-rule测试自定义配置中的rule设置。
二、使用
官方文档:https://elastalert.readthedocs.io
规则文档:https://elastalert.readthedocs.io/en/latest/ruletypes.html
1、主配置文件
首先是主配置文件的模板为config.yaml.example,生成全局配置
vim config.yaml
# 用来加载rule的目录,默认是example_rules
rules_folder: rules
# 用来设置定时向elasticsearch发送请求,也就是告警执行的频率
run_every:
seconds: 30
# 用来设置请求里时间字段的范围
buffer_time:
seconds: 30
# elasticsearch的host地址,端口
es_host: node01
es_port: 9200
# elastalert产生的日志在elasticsearch中的创建的索引
writeback_index: elastalert_status
writeback_alias: elastalert_alerts
# 失败重试的时间限制
alert_time_limit:
days: 2
2、创建告警索引
执行elastalert-create-index命令在ES创建索引,这不是必须的步骤,但是强烈建议创建。因为对于审计和测试很有用,并且重启ES不影响计数和发送alert。
$ elastalert-create-index
Elastic Version: 7.3.2
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!
看到这个输出,就说明创建成功了,也可以请求一下看看:
$ curl 127.0.0.1:9200/_cat/indices?v
health status index uuid pri rep docs.count docs.deleted store.size pri.store.size
green open elastalert_status_status lh8LL4iCQeSn0afyzxBX7w 1 1 0 0 460b 230b
green open elastalert_status i7B7IfCuSb2Sex8U5KoTZg 1 1 0 0 460b 230b
green open elastalert_status_past et2aF44VR4WQnxB8T7zD4Q 1 1 0 0 460b 230b
green open elastalert_status_silence lhXHEsuUQeGZaW3cRLp5pQ 1 1 0 0 460b 230b
green open elastalert_status_error zykwk4KtSyyOY7ckxQTrkA 1 1 0 0 460b 230b
3、Rule配置
所有的告警规则,通过在rule目下创建配置文件进行定义,这里简单创建一个来作为演示。
首先我已经在elk集群中配置了一个NGINX日志采集的流水线,现在去kibana中利用检索规则,过滤出我想要的告警内容,比如我想让状态码是404的请求,触发告警通知,就用如下语句进行查询:
response: 404
其中group是kafka里边定义的组,后边是状态码,还可以写更多条件进行匹配。
然后来到服务器添加一条规则:
vim nginx_404.yaml
name: Nginx_err
use_strftine_index: true
index: nginx_info*
type: any
aggregation:
seconds: 10
filter:
- query:
query_string:
query: "response: 404"
alert:
- "email"
email:
- "test@qq.com"
smtp_host: smtp.163.com
smtp_port: 25
smtp_auth_file: /opt/elastalert/smtp_auth_file.yaml
from_addr: test01@163.com
email_reply_to: teast02@163.com
注意里边在配置邮件通知的时候,还需要引用外部的一个文件,这个文件里用于存放对应邮箱的用户名密码,或者是邮箱邮箱的授权码
$ vim /opt/elastalert/smtp_auth_file.yaml
user: "test01@163.com"
password: "xxxxxxx"
4、规则测试
刚刚已经添加了一条规则,现在可以用自身的命令测试一下刚刚添加的规则。
$ elastalert-test-rule --config config.yaml nginx_404.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
.........
elastalert_status - {'rule_name': 'Nginx_err', 'endtime': datetime.datetime(2020, 1, 11, 7, 30, 59, 793352, tzinfo=tzutc()), 'starttime': datetime.datetime(2020, 1, 10, 7, 30, 59, 793352, tzinfo=tzutc()), 'matches': 0, 'hits': 0, '@timestamp': datetime.datetime(2020, 1, 11, 7, 31, 0, 76042, tzinfo=tzutc()), 'time_taken': 0.24003815650939941}
如果没有报错,则说明可用。
5、启动
启动方式有两种
(1)指定规则文件路径
python -m elastalert.elastalert --verbose --config config.yaml --rule rules/nginx_404.yaml
(2)在全局路径config.yaml下,配置规则存放在加载规则rules目录下
python -m elastalert.elastalert --verbose
6、验证
服务启动之后,日志能够很清晰看到整个过程,此时可以在刚刚的索引原点请求几个不存在的接口,造一些404状态,过一会儿应该可以看到日志中的说明,有告警发出,邮箱应该也能收到了。
三、优化
1、启动方式
上边的启动命令只是在前台启动,并不给力,可以使用nohup启动,或者是通过supervisord管理,会更加方便。 supervisord如何安装就不说了.
创建配置文件:
$cat /etc/supervisord.d/elastalert1.ini
[program:elastalert1]
directory=/data/elastalert1/
command=python -m elastalert.elastalert --verbose --config /data/elastalert1/config.yaml
process_name=elastalert1
autorestart=true
startsecs=15
stopsignal=INT
stopasgroup=true
killasgroup=true
redirect_stderr=true
stdout_logfile=/data/log/elastalert1.log
stdout_logfile_maxbytes=5MB
然后启动即可
supervisorctl update
supervisorctl start elastalert1
2、报警方式
elastalert的报警方式有很多种,像邮件、微信、钉钉、post等等,我们主要介绍以下几种常用的
(1)邮件报警
alert:
- "email"
email:
- "test@qq.com"
smtp_host: smtp.163.com
smtp_port: 25
smtp_auth_file: /opt/elastalert/smtp_auth_file.yaml
from_addr: test01@163.com
email_reply_to: teast02@163.com
修改/opt/elastalert/smtp_auth_file.yaml信息
(2)微信机器人报警
git clone https://github.com/anjia0532/elastalert-wechat-plugin.git
cp elastalert-wechat-plugin/elastalert_modules/* elastalert_modules/
添加报警方式
alert:
- "elastalert_modules.wechat_qiye_alert.WeChatAlerter"
#后台登陆后【设置】->【权限管理】->【普通管理组】->【创建并设置通讯录和应用权限】->【CorpID,Secret】
#设置微信企业号的appid
corp_id: xxx
#设置微信企业号的Secret
secret: xxx
#后台登陆后【应用中心】->【选择应用】->【应用id】
#设置微信企业号应用id
agent_id: xx
#部门id
party_id: xx
#用户微信号
user_id: xx
# 标签id,多个用 | 分隔
(3)钉钉报警方式
git clone https://github.com/xuyaoqiang/elastalert-dingtalk-plugin.git
cp elastalert-dingtalk-plugin/elastalert_modules/dingtalk_alert.py elastalert_modules/
添加报警方式
alert:
- "elastalert_modules.dingtalk_alert.DingTalkAlerter"
dingtalk_webhook: "https://oapi.dingtalk.com/robot/send?access_token=fb6500f4c85b8cfe66fa9586870f3ce16c848eab1e1cb23110388d6d443f1e"
dingtalk_msgtype: text
3、报警频率
#限定时间内,发生事件次数
num_events: 3
#与上面参数结合使用,表示在2分钟内发生3次就报警
timeframe:
minutes: 2
4、避免重复告警
避免一定时间段中重复告警,可以配置realert
和exponential_realert
这两个选项:
# 5分钟内相同的报警不会重复发送
realert:
minutes: 5
# 指数级扩大 realert 时间,中间如果有报警,
# 则按照5->10->20->40->60不断增大报警时间到制定的最大时间,
# 如果之后报警减少,则会慢慢恢复原始realert时间
exponential_realert:
hours: 1
5、聚合相同告警
# 根据报警的内容将相同的报警按照 name 来聚合
aggregation_key: name
# 聚合报警的内容,只展示 name 与 message
summary_table_fields:
- name
- message
6、告警内容格式化
可以自定义告警内容,内部是使用Python
的format
来实现的。
alert_subject: "Error {1} @{2}"
alert_subject_args:
- name
- "@timestamp"
alert_text_type: alert_text_only
alert_text: |
> Name: {1}
> Message: {2}
> Host: {3} ({4})
alert_text_args:
- name
- message
- hostname
- host
最后,整理了比较全的配置文件
name: test_err
use_strftine_index: true
index: filebeat-7.3.2-*
type: any
#将多个匹配项汇总到一个警报中。每次找到匹配项时,ElastAlert将等待该aggregation时间段,并将特定规则在该时间段内发生的所有匹配项一起发送。
aggregation:
seconds: 10
#限定时间内,发生事件次数
num_events: 3
#与上面参数结合使用,在几分钟内
timeframe:
minutes: 2
realert:
# 5分钟内相同的报警不会重复发送
minutes: 5
# 指数级扩大 realert 时间,中间如果有报警,
# 则按照5->10->20->40->60不断增大报警时间到制定的最大时间,
# 如果之后报警减少,则会慢慢恢复原始realert时间
exponential_realert:
hours: 1
filter:
- query:
query_string:
query: "404"
alert:
- "email"
#在邮件正文会显示你定义的alert_text
alert_text: "You have a err message!"
#用户认证文件,需要user和password两个属性
smtp_host: smtp.163.com
smtp_port: 25
smtp_auth_file: /opt/elastalert/smtp_auth_file.yaml
#从哪个邮箱发送
from_addr: test@163.com
#回复给那个邮箱
email_reply_to: test@163.com
email:
#接收报警邮件的邮箱
- "test04@163.com"
四、示例
1、监控日志Web攻击行为
1.1 修改nginx日志格式
log_format logstash_json '{"time": "$time_local", '
'"remote_addr": "$remote_addr", '
'"remote_user": "$remote_user", '
'"request": "$request", '
'"status": "$status", '
'"body_bytes_sent": "$body_bytes_sent", '
'"http_referer": "$http_referer", '
'"http_user_agent": "$http_user_agent", '
'"http_x_forwarded_for": "$http_x_forwarded_for", '
'"request_time": "$request_time", '
'"request_length": "$request_length", '
'"host": "$http_host"}';
1.2 编写监控规则
name: web attack
realert:
minutes: 5
index: logstash-*
type: frequency
num_events: 10
timeframe:
minutes: 1
query_key:
- name
realert:
minutes: 5
exponential_realert:
hours: 1
filter:
- query_string:
# sql insert xss detect
query: "request: select.+(from|limit) OR request: union(.*?)select OR request: into.+(dump|out)file OR
request: (base64_decode|sleep|benchmark|and.+1=1|and.+1=2|or%20|exec|information_schema|where%20|union%20|%2ctable_name%20|cmdshell|table_schema) OR
request: (iframe|script|body|img|layer|div|meta|style|base|object|input|onmouseover|onerror|onload) OR
request: .+etc.+passwd OR http_user_agent:(HTTrack|harvest|audit|dirbuster|pangolin|nmap|sqln|-scan|hydra|Parser|libwww|BBBike|sqlmap|w3af|owasp|Nikto|fimap|havij|PycURL|zmeu|BabyKrokodil|netsparker|httperf|bench) OR
status: (400|404|500|501)
NOT (request:_health.html OR remote_addr:222.222.222.222 )
"
smtp_host: smtp.qiye.163.com
smtp_port: 25
smtp_auth_file: /opt/elastalert/smtp_auth_file.yaml
email_reply_to: xxx@163.com
from_addr: xxx@163.com
alert:
- "email"
email:
- "shystartree@163.com"
alert_subject: "web attack may be by {0} at @{1}"
alert_subject_args:
- remote_addr
- time
alert_text_type: alert_text_only
alert_text: |
你好,服务器({})可能正在受到web攻击,请采取手段阻止!!!!
### 截止发邮件前匹配到的请求数:{}
> 发生时间: {}
> timestamp:{}
> attacker's ip: {}
> request: {}
> status:{}
> UA头:{}
>>> 参考来源:{}
alert_text_args:
- host
- num_hits
- time
- "@timestamp"
- remote_addr
- request
- status
- http_user_agent
- source
2、五分钟内流量总和超过200M就发邮件
run_every:
minutes: 5
name: flow
type: metric_aggregation
index: nginx_info
buffer_time:
minutes: 5
metric_agg_key: body_bytes_sent
metric_agg_type: sum
max_threshold: 209715200
use_run_every_query_size: true
alert_text_type: alert_text_only
alert_subject: "Alter 最近五分钟流量超200M,请注意!!!"
alert_text: |
最近五分钟总流量: {0} B
kibana url: http://xxxxx
alert_text_args:
- metric_body_bytes_sent_sum
smtp_host: smtp.qq.com
smtp_port: 25
smtp_auth_file: /opt/elastalert/smtp_auth_file.yaml
from_addr: "xxxx@qq.com"
alert:
- "email"
email:
- "xxxx@qq.com"
3、对后端请求超过3秒的发送邮件
es_host: 192.168.20.6
es_port: 9200
run_every:
seconds: 30
name: xxx_reponse_time
index: n-xxx-*
type: whitelist
compare_key: "request"
ignore_null: true
whitelist:
- /index.html
- /siteapp/ecsAuthentication/hasAuthentication
type: frequency
num_events: 1
timeframe:
seconds: 30
filter:
- query_string:
query: "upstream_response_time: >3 "
alert_text_type: alert_text_only
alert_subject: "Alter {0} 接口后端处理超过3秒!!!"
alert_subject_args:
- _index
html_table_title: "<h2>This is a heading</h2>"
alert_text: |
timestamp: {0}
request_method: {1}
request: {2}
request_body: {3}
request_time: {4} s
upstream_response_time: {5} s
body_bytes_sent: {6} B
status: {7}
remote_addr: {8}
http_x_forwarded_for: {9}
upstream_addr: {10}
agent: {11}
alert_text_args:
- timestamp
- request_method
- request
- request_body
- request_time
- upstream_response_time
- body_bytes_sent
- status
- remote_addr
- http_x_forwarded_for
- upstream_addr
- agent
smtp_host: smtp.qq.com
smtp_port: 25
smtp_auth_file: /opt/elastalert/rule_templates/smtp_auth_file.yaml
from_addr: "xxx@qq.com"
alert:
- "email"
email:
- "xxxxx@qq.com"