Prometheus学习笔记–Prometheus监控之elasticsearch集群 |坐而言不如起而行！ 二丫讲...

prometheus监控es，同样采用exporter的方案。

• 项目地址：
  • elasticsearch_exporter：https://github.com/justwatchcom/elasticsearch_exporter

1、安装部署

现有es三节点的集群，环境大概如下：

主机 组件 10.3.6.30–es-node1 es，nginx 10.3.6.125–es-node2 es 10.3.6.124–es-node3 es,kibana

接着分别在如上三台主机上进行如下配置：

1. wget https://github.com/justwatchcom/elasticsearch_exporter/releases/download/v1.1.0/elasticsearch_exporter-1.1.0.linux-amd64.tar.gz
2. tar xf elasticsearch_exporter-1.1.0.linux-amd64.tar.gz
3. mv elasticsearch_exporter-1.1.0.linux-amd64 /usr/local/elasticsearch_exporter

启动监控客户端：

1. nohup ./elasticsearch_exporter --web.listen-address ":9109" --es.uri http://10.3.6.30:9200 &

使用systemd管理：

1. cat /lib/systemd/system/es_exporter.service
2. [Unit]
3. Description=The es_exporter
4. After=network.target
5. [Service]
6. Type=simple
7. User=prometheus
8. ExecStart=/usr/local/elasticsearch_exporter/elasticsearch_exporter --web.listen-address ":9308" --es.uri http://127.0.0.1:9200
9. Restart=on-failure
10. [Install]
11. WantedBy=multi-user.target

1. systemctl daemon-reload
2. systemctl start es_exporter

查看metrics：

1. curl 127.0.0.1:9109/metrics

2，配置 prometheus.yml 添加监控目标

1. $ vim /usr/local/prometheus/prometheus.yml
2. - job_name: 'elasticsearch'
3. scrape_interval: 60s
4. scrape_timeout: 30s
5. metrics_path: "/metrics"
6. static_configs:
7. - targets:
8. - '10.3.0.41:9109'
9. labels:
10. service: elasticsearch

重启服务。

1. $ systemctl restart prometheus

或者通过命令热加载：

1. curl -XPOST localhost:9090/-/reload

5，配置 Grafana 的模板

模板通过json文件进行导入，文件就在解压的包内。

参考地址：https://shenshengkun.github.io/posts/550bdf86.html

或者通过如下ID进行导入：2322以及其他。

6，开启认证的启动方式

如果es开启了认证，那么启动的时候需要将用户名密码加载进去：

1. lasticsearch_exporter --web.listen-address ":9308" --es.uri http://username:[email protected]:9200 &

其中使用的是monitoring的用户密码。

当然，除去这种命令行的启动方式之外，还可以像上边一样，基于systemd进行管理，只需将认证的参数信息写入到如下内容当中：

1. $ cat /etc/default/elasticsearch_exporter
2. EXPORTER_ARGS="--es.uri=http://username:[email protected]:9200"

接着将启动配置文件封装成如下脚本：

1. $ cat /etc/init.d/elasticsearch_exporter
2. #!/bin/sh
3. # chkconfig: 2345 60 20
4. # description: elasticsearch_exporter
5. NAME=elasticsearch_exporter
6. SCRIPT="/usr/bin/${NAME}"
7. PIDFILE="/var/run/${NAME}.pid"
8. LOGFILE="/var/log/${NAME}.log"
9. ENVFILE="/etc/default/${NAME}"
10. USER="root"
11. URL='http://192.10.10.1'
12. EXPORTER_NAME=$NAME
13. EXPORTER_PORT="9114"
14. #获取本机ip
15. IP=$(grep "IPADDR" /etc/sysconfig/network-scripts/ifcfg-eth0 | grep -Eo "([0-9]{1,3}\.){3}[0-9]{1,3}")
16. register_exporter() {
17. json_data='{"service_id":"'${EXPORTER_NAME}${IP//./}'","job":"'${EXPORTER_NAME}'","ip":"'${IP}'","port":"'$EXPORTER_PORT'","tags":"","meta": {"hostname": "'$(hostname)'"}}'
18. curl --connect-timeout 2 -s -X POST -H "Content-type: application/json" -d "${json_data}" $URL 2>&1 > /dev/null
19. }
20. start() {
21. if [ -f "${PIDFILE}" ] && kill -0 $(cat "${PIDFILE}") &> /dev/null; then
22. echo "${NAME} already running with PID $(cat ${PIDFILE})" >&2
23. return 1
24. fi
25. echo "Starting ${NAME}" >&2
26. . "${ENVFILE}"
27. CMD="${SCRIPT} --web.listen-address=${IP}:${EXPORTER_PORT} --log.level=error ${EXPORTER_ARGS}"
28. su - "${USER}" -c "${CMD} &> ${LOGFILE} & echo \$! > ${PIDFILE}"
29. # echo "${NAME} started with PID $(cat ${PIDFILE})" >&2
30. sleep 1
31. if [ -f "${PIDFILE}" ] && kill -0 $(cat "${PIDFILE}") &> /dev/null; then
32. echo "${NAME} started successfully." >&2
33. register_exporter
34. else
35. echo "${NAME} was not started OK"
36. return 1
37. fi
38. }
39. stop() {
40. if [ ! -f "$PIDFILE" ] || ! kill -0 $(cat "$PIDFILE") &> /dev/null; then
41. echo "${NAME} not running" >&2
42. return 1
43. fi
44. echo "Stopping ${NAME}..." >&2
45. kill -15 $(cat "$PIDFILE")
46. rm -f "$PIDFILE"
47. echo "${NAME} stopped" >&2
48. }
49. status() {
50. if [ ! -f "$PIDFILE" ] || ! kill -0 $(cat "$PIDFILE") &> /dev/null; then
51. echo "${NAME} is not running" >&2
52. else
53. echo "${NAME} is running" >&2
54. fi
55. }
56. uninstall() {
57. echo -n "Are you really sure you want to uninstall ${NAME}? That cannot be undone. [yes|No] "
58. local SURE
59. read SURE
60. if [ "$SURE" = "yes" ]; then
61. stop
62. rm -f "$PIDFILE"
63. echo "Notice: log file is not be removed: '$LOGFILE'" >&2
64. update-rc.d -f <NAME> remove
65. rm -fv "$0"
66. fi
67. }
68. case "$1" in
69. start)
70. start
71. ;;
72. stop)
73. stop
74. ;;
75. uninstall)
76. uninstall
77. ;;
78. restart)
79. stop
80. start
81. ;;
82. status)
83. status
84. ;;
85. register)
86. register_exporter
87. ;;
88. *)
89. echo "Usage: $0 {start|stop|restart|status|register|uninstall}"
90. esac

此处服务启动之后将会自动注册到统一的注册中心去，而不必再手动添加配置。

---

---