Prometheus 应用
安装
Prometheus的单机安装比较简单，这里采用的是单机进行安装。Prometheus的相关插件下载地址
1 | # 下载Prometheus |
创建prometheus启动文件
1 | cat >/usr/lib/systemd/system/prometheus.service<<EOF |
部署node_exporter
主要用来监控服务器的基础信息，如: cpu、内存、磁盘、网卡。
1 | # 下载node_exporter |
设置node_exporter开机启动
1 | cat >/usr/lib/systemd/system/node_exporter.service<<EOF |
安装mysql_exporter
主要监控mysql数据库的信息
1 | # 下载mysql_exporter |
创建msql的连接权限
mysqld_exporter需要连接Mysql，首先为它创建用户并赋予所需要的权限：
1 | GRANT REPLICATION CLIENT, PROCESS ON . TO 'exporter'@'localhost' identified by '123456'; |
设置mysql_exporter开启启动
1 | cat >/usr/lib/systemd/system/mysql_exporter.service<<EOF |
使用granafa给 MySQLD_Exporter添加监控图表:
- 主从主群监控(模板7371)：
- 相关mysql 状态监控7362：
- 缓冲池状态7365
Prometheus基于文件的动态加载
基于文件的服务发现是最通用的方式。这种方式不需要依赖于任何的平台或者第三方服务。对于Prometheus而言也不可能支持所有的平台或者环境。通过基于文件的服务发现方式下，Prometheus会定时从文件中读取最新的Target信息，可以通过任意的方式将监控Target的信息写入即可。
Prometheus 可以通过JSON或者YAML格式的文件，定义所有的监控目标。下面我是通过yaml的文件格式来进行配置监控。在添加实例的时候添加了一些额外的标签信息。如: env、service、group等，实例中采集到的样本信息将包含这些标签信息，从而可以通过该标签按照环境对数据进行统计。
修改prometheus.yml
1 | cat prometheus.yml |
scrape_configs 这里我定义了三组，分别是监控docker、vm、mysql的，每一个组对应一个yml文件。对应的服务到对应的文件进行增加即可。还可以增加(zk、es、ng、redis)等服务。
创建被扫描的文件
1 | mkdir -p /opt/prometheus/sd_config && cd /opt/prometheus/sd_config |
- docker.yml
1 | - labels: |
- vm.yml
1 | - labels: |
- mysql.yml
1 | - labels: |
在Prometheus UI的Targets下就可以看到当前定义的yml文件中动态获取到实例信息以及监控任务的采集状态，同时在Labels列下会包含用户添加的自定义标签:
在Prometheus UI的service-discovery下可以看到我们定义的job类型
alertmanager 部署
1 | wget https://github.com/prometheus/alertmanager/releases/download/v0.20.0/alertmanager-0.20.0.linux-amd64.tar.gz |
设置alertmanager开机启动
1 | cat >/usr/lib/systemd/system/alertmanager.service<<EOF |
修改prometheus配置用于加载alertmanager和alertmanager rules
1 | cat prometheus.yml |
新建rules规则
node 规则
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34mkdir -p /opt/prometheus/rules
cat >node.rules<<EOF
groups:
- name: kxl_Instances
rules:
- alert: InstanceDown
expr: up == 0
for: 5m
labels:
severity: page
# Prometheus templates apply here in the annotation and label fields of the alert.
annotations:
description: '{{ $labels.instance }} of job {{ $labels.job }} has been down for more than 5 minutes.'
summary: 'Instance {{ $labels.instance }} down'
- alert: 内存使用率过高
expr: 100-(node_memory_Buffers_bytes+node_memory_Cached_bytes+node_memory_MemFree_bytes)/node_memory_MemTotal_bytes*100 > 30
for: 1m
labels:
severity: warning
annotations:
summary: "Instance {{ $labels.instance }} 内存使用率过高"
description: "{{ $labels.instance }} of job {{$labels.job}}内存使用率超过80%,当前使用率[{{ $value }}]."
- alert: cpu使用率过高
expr: 100-avg(irate(node_cpu_seconds_total{mode="idle"}[5m])) by(instance)*100 > 0
for: 1m
labels:
severity: warning
annotations:
summary: "Instance {{ $labels.instance }} cpu使用率过高"
description: "{{ $labels.instance }} of job {{$labels.job}}cpu使用率超过80%,当前使用率[{{ $value }}]."
EOFmysql 规则
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198cat > mysql.rules <<EOF
groups:
- name: MySQLStatsAlert
rules:
- alert: MySQL is down
expr: mysql_up == 0
for: 1m
labels:
severity: critical
annotations:
summary: "Instance {{ $labels.instance }} MySQL is down"
description: "MySQL database is down. This requires immediate action!"
- alert: open files high
expr: mysql_global_status_innodb_num_open_files > (mysql_global_variables_open_files_limit) * 0.75
for: 1m
labels:
severity: warning
annotations:
summary: "Instance {{ $labels.instance }} open files high"
description: "Open files is high. Please consider increasing open_files_limit."
- alert: Read buffer size is bigger than max. allowed packet size
expr: mysql_global_variables_read_buffer_size > mysql_global_variables_slave_max_allowed_packet
for: 1m
labels:
severity: warning
annotations:
summary: "Instance {{ $labels.instance }} Read buffer size is bigger than max. allowed packet size"
description: "Read buffer size (read_buffer_size) is bigger than max. allowed packet size (max_allowed_packet).This can break your replication."
- alert: Sort buffer possibly missconfigured
expr: mysql_global_variables_innodb_sort_buffer_size <256*1024 or mysql_global_variables_read_buffer_size > 4*1024*1024
for: 1m
labels:
severity: warning
annotations:
summary: "Instance {{ $labels.instance }} Sort buffer possibly missconfigured"
description: "Sort buffer size is either too big or too small. A good value for sort_buffer_size is between 256k and 4M."
- alert: Thread stack size is too small
expr: mysql_global_variables_thread_stack <196608
for: 1m
labels:
severity: warning
annotations:
summary: "Instance {{ $labels.instance }} Thread stack size is too small"
description: "Thread stack size is too small. This can cause problems when you use Stored Language constructs for example. A typical is 256k for thread_stack_size."
- alert: Used more than 80% of max connections limited
expr: mysql_global_status_max_used_connections > mysql_global_variables_max_connections * 0.8
for: 1m
labels:
severity: warning
annotations:
summary: "Instance {{ $labels.instance }} Used more than 80% of max connections limited"
description: "Used more than 80% of max connections limited"
- alert: InnoDB Force Recovery is enabled
expr: mysql_global_variables_innodb_force_recovery != 0
for: 1m
labels:
severity: warning
annotations:
summary: "Instance {{ $labels.instance }} InnoDB Force Recovery is enabled"
description: "InnoDB Force Recovery is enabled. This mode should be used for data recovery purposes only. It prohibits writing to the data."
- alert: InnoDB Log File size is too small
expr: mysql_global_variables_innodb_log_file_size < 16777216
for: 1m
labels:
severity: warning
annotations:
summary: "Instance {{ $labels.instance }} InnoDB Log File size is too small"
description: "The InnoDB Log File size is possibly too small. Choosing a small InnoDB Log File size can have significant performance impacts."
- alert: InnoDB Flush Log at Transaction Commit
expr: mysql_global_variables_innodb_flush_log_at_trx_commit != 1
for: 1m
labels:
severity: warning
annotations:
summary: "Instance {{ $labels.instance }} InnoDB Flush Log at Transaction Commit"
description: "InnoDB Flush Log at Transaction Commit is set to a values != 1. This can lead to a loss of commited transactions in case of a power failure."
- alert: Table definition cache too small
expr: mysql_global_status_open_table_definitions > mysql_global_variables_table_definition_cache
for: 1m
labels:
severity: page
annotations:
summary: "Instance {{ $labels.instance }} Table definition cache too small"
description: "Your Table Definition Cache is possibly too small. If it is much too small this can have significant performance impacts!"
- alert: Table open cache too small
expr: mysql_global_status_open_tables >mysql_global_variables_table_open_cache * 99/100
for: 1m
labels:
severity: page
annotations:
summary: "Instance {{ $labels.instance }} Table open cache too small"
description: "Your Table Open Cache is possibly too small (old name Table Cache). If it is much too small this can have significant performance impacts!"
- alert: Thread stack size is possibly too small
expr: mysql_global_variables_thread_stack < 262144
for: 1m
labels:
severity: page
annotations:
summary: "Instance {{ $labels.instance }} Thread stack size is possibly too small"
description: "Thread stack size is possibly too small. This can cause problems when you use Stored Language constructs for example. A typical is 256k for thread_stack_size."
- alert: InnoDB Buffer Pool Instances is too small
expr: mysql_global_variables_innodb_buffer_pool_instances == 1
for: 1m
labels:
severity: page
annotations:
summary: "Instance {{ $labels.instance }} InnoDB Buffer Pool Instances is too small"
description: "If you are using MySQL 5.5 and higher you should use several InnoDB Buffer Pool Instances for performance reasons. Some rules are: InnoDB Buffer Pool Instance should be at least 1 Gbyte in size. InnoDB Buffer Pool Instances you can set equal to the number of cores of your machine."
- alert: InnoDB Plugin is enabled
expr: mysql_global_variables_ignore_builtin_innodb == 1
for: 1m
labels:
severity: page
annotations:
summary: "Instance {{ $labels.instance }} InnoDB Plugin is enabled"
description: "InnoDB Plugin is enabled"
- alert: Binary Log is disabled
expr: mysql_global_variables_log_bin != 1
for: 1m
labels:
severity: warning
annotations:
summary: "Instance {{ $labels.instance }} Binary Log is disabled"
description: "Binary Log is disabled. This prohibits you to do Point in Time Recovery (PiTR)."
- alert: Binlog Cache size too small
expr: mysql_global_variables_binlog_cache_size < 1048576
for: 1m
labels:
severity: page
annotations:
env: "{{ $labels.env }}"
summary: "Instance {{ $labels.instance }} Binlog Cache size too small"
description: "Binlog Cache size is possibly to small. A value of 1 Mbyte or higher is OK."
- alert: Binlog Statement Cache size too small
expr: mysql_global_variables_binlog_stmt_cache_size <1048576 and mysql_global_variables_binlog_stmt_cache_size > 0
for: 1m
labels:
severity: page
annotations:
summary: "Instance {{ $labels.instance }} Binlog Statement Cache size too small"
description: "Binlog Statement Cache size is possibly to small. A value of 1 Mbyte or higher is typically OK."
- alert: Binlog Transaction Cache size too small
expr: mysql_global_variables_binlog_cache_size <1048576
for: 1m
labels:
severity: page
annotations:
summary: "Instance {{ $labels.instance }} Binlog Transaction Cache size too small"
description: "Binlog Transaction Cache size is possibly to small. A value of 1 Mbyte or higher is typically OK."
- alert: Sync Binlog is enabled
expr: mysql_global_variables_sync_binlog == 1
for: 1m
labels:
severity: page
annotations:
summary: "Instance {{ $labels.instance }} Sync Binlog is enabled"
description: "Sync Binlog is enabled. This leads to higher data security but on the cost of write performance."
- alert: IO thread stopped
expr: mysql_slave_status_slave_io_running != 1
for: 1m
labels:
severity: critical
annotations:
summary: "Instance {{ $labels.instance }} IO thread stopped"
description: "IO thread has stopped. This is usually because it cannot connect to the Master any more."
- alert: SQL thread stopped
expr: mysql_slave_status_slave_sql_running == 0
for: 1m
labels:
severity: critical
annotations:
summary: "Instance {{ $labels.instance }} SQL thread stopped"
description: "SQL thread has stopped. This is usually because it cannot apply a SQL statement received from the master."
- alert: SQL thread stopped
expr: mysql_slave_status_slave_sql_running != 1
for: 1m
labels:
severity: critical
annotations:
summary: "Instance {{ $labels.instance }} Sync Binlog is enabled"
description: "SQL thread has stopped. This is usually because it cannot apply a SQL statement received from the master."
- alert: Slave lagging behind Master
expr: rate(mysql_slave_status_seconds_behind_master[1m]) >30
for: 1m
labels:
severity: warning
annotations:
summary: "Instance {{ $labels.instance }} Slave lagging behind Master"
description: "Slave is lagging behind Master. Please check if Slave threads are running and if there are some performance issues!"
- alert: Slave is NOT read only(Please ignore this warning indicator.)
expr: mysql_global_variables_read_only != 0
for: 1m
labels:
severity: page
annotations:
summary: "Instance {{ $labels.instance }} Slave is NOT read only"
description: "Slave is NOT set to read only. You can accidentally manipulate data on the slave and get inconsistencies..."
EOF
配置告警策略
1 | cat alertmanager.yml |
Alertmanager主要负责对Prometheus产生的告警进行统一处理，因此在Alertmanager配置中一般会包含以下几个主要部分：
- 全局配置（global）：用于定义一些全局的公共参数，如全局的SMTP配置，Slack配置等内容；
- 模板（templates）：用于定义告警通知时的模板，如HTML模板，邮件模板等；
- 告警路由（route）：根据标签匹配，确定当前告警应该如何处理；
- 接收人（receivers）：接收人是一个抽象的概念，它可以是一个邮箱也可以是微信，Slack或者Webhook等，接收人一般配合告警路由使用；
- 抑制规则（inhibit_rules）：合理设置抑制规则可以减少垃圾告警的产生
.tmpl模板的配置
1 | # 创建.tmpl模版存放目录 |
企业微信截图
