elasticsearch之watcher插件安装文档

elasticsearch之watcher插件安装文档

watcher插件可以提供elasticsearch集群的监控提醒功能。
You need to install the License and Watcher plugins on each node in your cluster.
你需要在你的集群中所有节点安装license和watcher插件

安装watcher插件
1：安装license
bin/plugin install license

2：安装watcher插件
bin/plugin install watcher
3：在安装的时候需要确认一下
Continue with installation? [y/N]y
Specify the –batch option when running the install command to automatically grant these permissions and bypass this install prompt.
或者指定在安装插件命令后面添加 –batch，忽略确认直接安装。
4：【默认这一步不需要操作，因为默认情况下es并没有禁用自动创建索引库功能】If you have disabled automatic index creation in Elasticsearch, configure action.auto_create_index in elasticsearch.yml to allow watcher to create the .watches, .triggered_watches, and .watcher-history* indices:
如果你在es中禁用了自动创建索引库的功能。那么你需要在elasticsearch.yml中配置action.auto_create_index 以便允许watcher去创建.watches, .triggered_watches, and .watcher-history* 索引库。

action.auto_create_index: .watches,.triggered_watches,.watcher-history*

If you have Shield installed, you must also allow Shield to create the .security index.
如果你已经安装了shield，你必须允许shield创建.security索引库

5：启动es
bin/elasticsearch

6：验证
curl -XGET ‘http://localhost:9200/_watcher/stats?pretty’

{
“watcher_state”: “started”,
“watch_count”: 0,
“execution_thread_pool”: {
“queue_size”: 0,
“max_size”: 0
}
}

监控功能：
1：监控错误日志
Watch Log Data for Errors
https://www.elastic.co/guide/en/watcher/current/watch-log-data.html
2：监控集群监控状态
Watch Your Cluster Health
https://www.elastic.co/guide/en/watcher/current/watch-cluster-status.html

下面以监控集群状态为例演示：
1：创建一个每10秒运行一次的调度器
curl -XPUT ‘http://192.168.80.100:9200/_watcher/watch/cluster_health_watch’ -d ‘{
“trigger” : {
“schedule” : { “interval” : “10s” }
}
}’

2：给这个调度器增加一个input，把集群的健康状态添加到watcher中
curl -XPUT ‘http://192.168.80.100:9200/_watcher/watch/cluster_health_watch’ -d ‘{
“trigger” : {
“schedule” : { “interval” : “10s” }
},
“input” : {
“http” : {
“request” : {
“host” : “192.168.80.100”,
“port” : 9200,
“path” : “/_cluster/health”
}
}
}
}’

3：给这个wacher添加一个条件,判断集群的状态是否为red
curl -XPUT ‘http://192.168.80.100:9200/_watcher/watch/cluster_health_watch’ -d ‘{
“trigger” : {
“schedule” : { “interval” : “10s” }
},
“input” : {
“http” : {
“request” : {
“host” : “192.168.80.100”,
“port” : 9200,
“path” : “/_cluster/health”
}
}
},
“condition” : {
“compare” : {
“ctx.payload.status” : { “eq” : “red” }
}
}
}’

4：给watcher设置触发任务，当满足条件的时候，执行。
一般都是发送邮件。
所以需要先配置邮箱信息。修改es的配置文件。
vi conf/elasticsearch.yml
watcher.actions.email.service.account:
work:
profile: qqmail
smtp:
auth: true
starttls.enable: true
host: smtp.qq.com
port: 587
user: [e-mail]
password: 12345678
添加触发任务【手动把某一个索引库的分片删除一部分可以模拟分片丢失，集群状态会变为red，此任务就会被触发，发送邮件。】
curl -XPUT ‘http://192.168.80.100:9200/_watcher/watch/cluster_health_watch’ -d ‘{
“trigger” : {
“schedule” : { “interval” : “10s” }
},
“input” : {
“http” : {
“request” : {
“host” : “192.168.80.100”,
“port” : 9200,
“path” : “/_cluster/health”
}
}
},
“condition” : {
“compare” : {
“ctx.payload.status” : { “eq” : “yellow” }
}
},
“actions” : {
“send_email” : {
“email” : {
“to” : “[e-mail]”,
“subject” : “Cluster Status Warning”,
“body” : “Cluster status is RED”
}
}
}
}’

验证集群状态，如果集群状态为red，则刚才配置的watcher任务就会执行，发送邮件提醒。

curl -XGET ‘http://192.168.80.100:9200/_cluster/health?pretty’
5：实验结束，可以移除掉这个watcher
curl -XDELETE ‘http://192.168.80.100:9200/_watcher/watch/cluster_health_watch’

卸载watcher
To uninstall Watcher:

1：停止Elasticsearch.
2：从Elasticsearch中移除watcher:
bin/plugin remove watcher

3：启动Elasticsearch.

链接：http://www.bigdata1024.com/?p=43