1-监控界面sql保存

1, BufferSize_machine

1), template

主要用来监控buffersize的状态的

name: 模块名字, 用于后续调取使用,

label: 模块显示名字, 在页面显示的

includeAll: 是否包含 all 按钮

query: 查询的sql语句, 由于模版一致, 所以后续只保留sql

classify:

select distinct(classify) from host_dict

model_name

select distinct(model_name) from host_dict where classify in ($classify)

2) KafkaSinkNetword

SELECT
  UNIX_TIMESTAMP(time) as time_sec,
  value as value,
  host as metric
FROM jmx_status as jmx
left join host_dict as hd 
  on hd.innet_ip = jmx.host
WHERE $__timeFilter(time)
  AND attribute = "BufferSize"
  AND object = 'KafkaSinkNetwork'
  AND hd.classify in ($classify)
  and hd.model_name in ($model_name)
ORDER BY time ASC, host

3), KafkaSinkFile

SELECT
  UNIX_TIMESTAMP(time) as time_sec,
  value as value,
  host as metric
FROM jmx_status as jmx
left join host_dict as hd 
  on hd.innet_ip = jmx.host
WHERE $__timeFilter(time)
  AND attribute = "BufferSize"
  AND object = 'KafkaSinkFile'
  AND hd.classify in ($classify)
  and hd.model_name in ($model_name)
ORDER BY time ASC, host

4), FileSink

SELECT
  UNIX_TIMESTAMP(time) as time_sec,
  value as value,
  host as metric
FROM jmx_status as jmx
left join host_dict as hd 
  on hd.innet_ip = jmx.host
WHERE $__timeFilter(time)
  AND attribute = "BufferSize"
  AND object = 'FileSink'
  AND hd.classify in ($classify)
  and hd.model_name in ($model_name)
ORDER BY time ASC, host

5), MessageCopy

SELECT
  UNIX_TIMESTAMP(time) as time_sec,
  value as value,
  host as metric
FROM jmx_status as jmx
left join host_dict as hd 
  on hd.innet_ip = jmx.host
WHERE $__timeFilter(time)
  AND attribute = "BufferSize"
  AND object = 'MessageCopy'
  AND hd.classify in ($classify)
  and hd.model_name in ($model_name)
ORDER BY time ASC, host

2, BufferSizeTopic

1) BufferSize长度

SELECT
  UNIX_TIMESTAMP(time) as time_sec,
  sum(value) as value,
  object as metric
FROM jmx_status
WHERE $__timeFilter(time) 
  AND attribute = 'BufferSize'
GROUP BY object, time
ORDER BY time ASC, metric

3, Metric_machine

1), template

Classify

select distinct(classify) from host_dict

model_name

select distinct(model_name) from host_dict where classify in ($classify)

source_file

select distinct(component) from topic_count where component like '%Source'

kafka_file

select distinct(component) from topic_count where component like 'Kafka%';

2), 接受消息总量

SELECT
  UNIX_TIMESTAMP(tc.time) as time_sec,
  tc.host as metric,
  SUM(out_num) as value
FROM topic_count as tc
left join topic_dict as td
  on tc.topic = td.topic
left join host_dict as hd
  on hd.innet_ip = tc.host
WHERE $__timeFilter(time) 
  AND tc.component in ($source_file)
  AND hd.classify in ($classify)
  and hd.model_name in ($model_name)
GROUP BY tc.host, tc.time
ORDER BY tc.time ASC, tc.host

3), 发送kafka消息总量

SELECT
  UNIX_TIMESTAMP(tc.time) as time_sec,
  tc.host as metric,
  SUM(in_num - out_num) as value
FROM topic_count as tc
left join topic_dict as td
  on tc.topic = td.topic
left join host_dict as hd
  on hd.innet_ip = tc.host
WHERE $__timeFilter(time) 
  AND tc.component in ($kafka_file)
  AND hd.classify in ($classify)
  and hd.model_name in ($model_name)
GROUP BY tc.host, tc.time
ORDER BY tc.time ASC, tc.host

4), 发送kafka消息失败量

SELECT
  UNIX_TIMESTAMP(tc.time) as time_sec,
  tc.host as metric,
  SUM(out_num) as value
FROM topic_count as tc
left join host_dict as hd
  on hd.innet_ip = tc.host
WHERE $__timeFilter(time) 
 AND (tc.component = 'KafkaSinkNetwork' OR tc.component = 'KafkaSinkFile')
  AND hd.classify in ($classify)
  and hd.model_name in ($model_name)
GROUP BY tc.host, tc.time
ORDER BY tc.time ASC, tc.host

4, Metric_topic

1), template

component_name

select distinct(component_name) from topic_dict

topic

select distinct(topic) from topic_dict where component_name in ($component_name)

source_file

select distinct(component) from topic_count where component like '%Source'

kafka_file

select distinct(component) from topic_count where component like 'Kafka%';

2), iris接收总量

SELECT
  UNIX_TIMESTAMP(tc.time) as time_sec,
  tc.topic as metric,
  SUM(out_num) as value
FROM topic_count as tc
left join topic_dict as td
  on tc.topic = td.topic
WHERE $__timeFilter(time) 
  AND tc.topic in ($topic)
  AND tc.component in ($source_file)
GROUP BY tc.topic, tc.time
ORDER BY tc.time ASC, tc.topic

3), 发送kafka消息总量

SELECT
  UNIX_TIMESTAMP(time) as time_sec,
  tc.topic as metric, 
  SUM(in_num - out_num) as value
FROM topic_count as tc
left join topic_dict as td
  on tc.topic = td.topic
WHERE $__timeFilter(time)
  AND tc.topic in ($topic)
  AND tc.component in ($kafka_file)
GROUP BY tc.topic, tc.time
ORDER BY tc.time ASC, tc.topic

4), 发送kafka消息失败量

SELECT
  UNIX_TIMESTAMP(time) as time_sec,
  sum(out_num) as value,
  topic as metric
FROM topic_count
WHERE $__timeFilter(time) 
  AND (component = 'KafkaSinkNetwork' OR component = 'KafkaSinkFile')
  AND topic in ($topic)
GROUP BY topic, time
ORDER BY time ASC, metric

5), 消息丢失数

SELECT
  UNIX_TIMESTAMP(time) as time_sec,
  (temp.value - SUM(tc.in_num - tc.out_num)) as value, 
  tc.topic as metric
FROM topic_count as tc
left join topic_dict as td
  on tc.topic = td.topic
right join (
    SELECT
      tc2.time as calen,
      tc2.topic, 
      SUM(out_num) as value
    FROM topic_count as tc2
    left join topic_dict as td2
      on tc2.topic = td2.topic
    WHERE $__timeFilter(tc2.time)
      AND tc2.topic in ($topic)
      AND tc2.component in ($source_file)
      AND tc2.component <> 'FileSource'
    GROUP BY tc2.topic, tc2.time
) as temp
  on temp.calen = tc.time
    and temp.topic = tc.topic
WHERE $__timeFilter(time)
  AND tc.topic in ($topic)
  AND tc.component in ($kafka_file)
GROUP BY tc.time, tc.topic
ORDER BY tc.topic, tc.time asc

6), 验平汇总, 此为表格

SELECT
  date_format(time, '%Y-%m-%d %H:%i:%s') as time,
  tc.topic, 
  temp.value as iris总接受量, 
  SUM(tc.in_num - tc.out_num) as kafka发送成功, 
  SUM(tc.out_num) as kafka发送失败, 
  (temp.value - SUM(tc.in_num - tc.out_num)) as 消息丢失数
FROM topic_count as tc
left join topic_dict as td
  on tc.topic = td.topic
right join (
    SELECT
      tc2.time as calen,
      tc2.topic, 
      SUM(out_num) as value
    FROM topic_count as tc2
    left join topic_dict as td2
      on tc2.topic = td2.topic
    WHERE $__timeFilter(tc2.time)
      AND tc2.topic in ($topic)
      AND tc2.component in ($source_file)
      AND tc2.component <> 'FileSource'
    GROUP BY tc2.topic, tc2.time
) as temp
  on temp.calen = tc.time
    and temp.topic = tc.topic
WHERE $__timeFilter(time)
  AND tc.topic in ($topic)
  AND tc.component in ($kafka_file)
GROUP BY tc.time, tc.topic
ORDER BY tc.topic, tc.time asc

5, QPS_Component

qps

SELECT
  UNIX_TIMESTAMP(time) as time_sec,
  sum(value) as value,
  object as metric
FROM jmx_status
WHERE $__timeFilter(time) 
  AND attribute = 'QPS'
GROUP BY object, time
ORDER BY time ASC, metric

6, QPS_machine

1), template

classify

select distinct(classify) from host_dict

model_name

select distinct(model_name) from host_dict where classify in ($classify)

2), topic_source

SELECT
  UNIX_TIMESTAMP(time) as time_sec,
  value as value,
  host as metric
FROM jmx_status as jmx
left join host_dict as hd 
  on hd.innet_ip = jmx.host
WHERE $__timeFilter(time)
  AND attribute = "QPS"
  AND object = 'TcpSource'
  AND hd.classify in ($classify)
  and hd.model_name in ($model_name)
ORDER BY time DESC, metric
limit 990000

3), js_source

SELECT
  UNIX_TIMESTAMP(time) as time_sec,
  value as value,
  host as metric
FROM jmx_status as jmx
left join host_dict as hd 
  on hd.innet_ip = jmx.host
WHERE $__timeFilter(time)
  AND attribute = "QPS"
  AND object = 'JsSource'
  AND hd.classify in ($classify)
  and hd.model_name in ($model_name)
ORDER BY time DESC
limit 990000

4), legency_source

SELECT
  UNIX_TIMESTAMP(time) as time_sec,
  value as value,
  host as metric
FROM jmx_status as jmx
left join host_dict as hd 
  on hd.innet_ip = jmx.host
WHERE object = 'LegacyJsSource'
  AND attribute = "QPS"
  AND $__timeFilter(time)
  AND hd.classify in ($classify)
  and hd.model_name in ($model_name)
ORDER BY time DESC
limit 990000

5), webSource

SELECT
  UNIX_TIMESTAMP(time) as time_sec,
  value as value,
  host as metric
FROM jmx_status as jmx
left join host_dict as hd 
  on hd.innet_ip = jmx.host
WHERE object = 'WebSource'
  AND attribute = "QPS"
  AND $__timeFilter(time)
  AND hd.classify in ($classify)
  and hd.model_name in ($model_name)
ORDER BY time DESC
limit 990000

6), zhixinSource

SELECT
  UNIX_TIMESTAMP(time) as time_sec,
  value as value,
  concat('KSF-', host) as metric
FROM jmx_status as jmx
left join host_dict as hd 
  on hd.innet_ip = jmx.host
WHERE object = 'ZhixinSource'
  AND attribute = "QPS"
  AND $__timeFilter(time)
  AND hd.classify in ($classify)
  and hd.model_name in ($model_name)
ORDER BY time DESC
limit 990000

7) cdn_httpsource

SELECT
  UNIX_TIMESTAMP(time) as time_sec,
  value as value,
  host as metric
FROM jmx_status as jmx
left join host_dict as hd 
  on hd.innet_ip = jmx.host
WHERE object = 'CdnHttpSource'
  AND attribute = "QPS"
  AND $__timeFilter(time)
  AND hd.classify in ($classify)
  and hd.model_name in ($model_name)
ORDER BY time DESC
limit 990000

8), qps_everyhost

SELECT
  UNIX_TIMESTAMP(time) as time_sec,
  sum(value) as value,
  host as metric
FROM jmx_status as jmx
left join host_dict as hd 
  on hd.innet_ip = jmx.host
WHERE $__timeFilter(time)
  AND attribute = "QPS"
  AND hd.classify in ($classify)
  and hd.model_name in ($model_name)
group by host, time
ORDER BY time DESC, metric
limit 990000

9), qps_hostnum

SELECT
  UNIX_TIMESTAMP(js.time) as time_sec,
  count(distinct(js.host)) as value, 
  hd.classify as metric
FROM jmx_status as js
left join host_dict as hd
  on js.host = hd.innet_ip
WHERE $__timeFilter(time)
  AND attribute = "QPS"
group by time, hd.classify
ORDER BY time DESC, metric
limit 990000

7, Resource_machine

1), template

classify

select distinct(classify) from host_dict

model_name

select distinct(model_name) from host_dict where classify in ($classify)

2), cpu_used

SELECT
  UNIX_TIMESTAMP(time) as time_sec,
  value as value,
  host as metric
FROM jmx_status as jmx
left join host_dict as hd 
  on hd.innet_ip = jmx.host
WHERE $__timeFilter(time)
  AND attribute = "SystemCpuLoad"
  AND hd.classify in ($classify)
  and hd.model_name in ($model_name)
ORDER BY time ASC, metric

3), memory_used

SELECT
  UNIX_TIMESTAMP(time) as time_sec,
  value as value,
  host as metric
FROM jmx_status as jmx
left join host_dict as hd 
  on hd.innet_ip = jmx.host
WHERE $__timeFilter(time)
  AND attribute = "HeapMemoryUsage.used"
  AND hd.classify in ($classify)
  and hd.model_name in ($model_name)
ORDER BY time ASC, metric

4), thread_count

SELECT
  UNIX_TIMESTAMP(time) as time_sec,
  value as value,
  host as metric
FROM jmx_status as jmx
left join host_dict as hd 
  on hd.innet_ip = jmx.host
WHERE $__timeFilter(time)
  AND attribute = "ThreadCount"
  AND hd.classify in ($classify)
  and hd.model_name in ($model_name)
ORDER BY time ASC, metric

5), openfile_script

SELECT
  UNIX_TIMESTAMP(time) as time_sec,
  value as value,
  host as metric
FROM jmx_status as jmx
left join host_dict as hd 
  on hd.innet_ip = jmx.host
WHERE $__timeFilter(time)
  AND attribute = "OpenFileDescriptorCount"
  AND hd.classify in ($classify)
  and hd.model_name in ($model_name)
ORDER BY time ASC, metric

8, alter 报警用, 需要配合设置

iris第一次尝试入kafka失败百分比

SELECT
  UNIX_TIMESTAMP(time) as time_sec,
  value as value,
  host as metric
FROM jmx_status as jmx
left join host_dict as hd 
  on hd.innet_ip = jmx.host
WHERE $__timeFilter(time)
  AND attribute = "OpenFileDescriptorCount"
  AND hd.classify in ($classify)
  and hd.model_name in ($model_name)
ORDER BY time ASC, metric

主要在alter标签中

然后在alter标签中进行配置

9, 需要用到的sql保存

1), host_dict

CREATE TABLE `host_dict` (
  `id` bigint(20) NOT NULL AUTO_INCREMENT COMMENT '主键id',
  `classify` varchar(20) DEFAULT NULL COMMENT '类型',
  `model_name` varchar(20) DEFAULT NULL COMMENT '模块名',
  `innet_ip` varchar(20) DEFAULT NULL COMMENT '内网ip',
  `outnet_ip` varchar(20) DEFAULT NULL COMMENT '外网ip',
  `cpu_core` int(11) DEFAULT NULL COMMENT 'cpu核心',
  `memory_size` int(11) DEFAULT NULL COMMENT '内存',
  `address` varchar(20) DEFAULT NULL COMMENT '机房',
  `status` varchar(20) DEFAULT NULL COMMENT '状态',
  `plan` varchar(20) DEFAULT NULL COMMENT '规划',
  PRIMARY KEY (`id`),
  KEY `classify` (`classify`,`model_name`,`innet_ip`),
  KEY `idx_innet_ip` (`innet_ip`) USING BTREE
) ENGINE=InnoDB AUTO_INCREMENT=49 DEFAULT CHARSET=utf8

2), jmx_status

CREATE TABLE `jmx_status` (
  `host` varchar(20) NOT NULL DEFAULT '',
  `time` timestamp NOT NULL DEFAULT CURRENT_TIMESTAMP ON UPDATE CURRENT_TIMESTAMP,
  `report` timestamp NOT NULL DEFAULT '0000-00-00 00:00:00',
  `object` varchar(60) NOT NULL DEFAULT '',
  `attribute` varchar(60) NOT NULL DEFAULT '',
  `value` double DEFAULT NULL,
  PRIMARY KEY (`host`,`time`,`object`,`attribute`),
  KEY `idx_host_time_attribute_object` (`host`,`time`,`attribute`,`object`) USING BTREE,
  KEY `idx_time_attribute` (`time`,`attribute`) USING BTREE
) ENGINE=InnoDB DEFAULT CHARSET=utf8
/*!50100 PARTITION BY RANGE (unix_timestamp(time))
(PARTITION p20180811 VALUES LESS THAN (1534003199) ENGINE = InnoDB,
 PARTITION p20180812 VALUES LESS THAN (1534089599) ENGINE = InnoDB,
 PARTITION p20180813 VALUES LESS THAN (1534175999) ENGINE = InnoDB,
 PARTITION p20180814 VALUES LESS THAN (1534262399) ENGINE = InnoDB,
 PARTITION p20180815 VALUES LESS THAN (1534348799) ENGINE = InnoDB,
 PARTITION p20180816 VALUES LESS THAN (1534435199) ENGINE = InnoDB,
 PARTITION p20180817 VALUES LESS THAN (1534521599) ENGINE = InnoDB,
 PARTITION p20180818 VALUES LESS THAN (1534607999) ENGINE = InnoDB,
 PARTITION p20180819 VALUES LESS THAN (1534694399) ENGINE = InnoDB,
 PARTITION p20180820 VALUES LESS THAN (1534780799) ENGINE = InnoDB,
 PARTITION p20180821 VALUES LESS THAN (1534867199) ENGINE = InnoDB,
 PARTITION p20180822 VALUES LESS THAN (1534953599) ENGINE = InnoDB) */

3), topic_count

CREATE TABLE `topic_count` (
  `host` varchar(20) NOT NULL DEFAULT '',
  `time` timestamp NOT NULL DEFAULT CURRENT_TIMESTAMP ON UPDATE CURRENT_TIMESTAMP,
  `report` timestamp NOT NULL DEFAULT '0000-00-00 00:00:00',
  `component` varchar(32) NOT NULL DEFAULT '',
  `topic` varchar(32) NOT NULL DEFAULT '',
  `in_num` bigint(20) DEFAULT NULL,
  `out_num` bigint(20) DEFAULT NULL,
  PRIMARY KEY (`host`,`time`,`topic`,`component`),
  KEY `component` (`component`,`topic`,`time`),
  KEY `idx_topic` (`topic`) USING BTREE,
  KEY `idx_time_topic` (`time`,`topic`) USING BTREE,
  KEY `idx_compnent` (`component`) USING BTREE
) ENGINE=InnoDB DEFAULT CHARSET=utf8

4), topic_dict

CREATE TABLE `topic_dict` (
  `id` bigint(20) NOT NULL AUTO_INCREMENT COMMENT '主键id',
  `model` varchar(20) DEFAULT NULL COMMENT '模式',
  `component` varchar(20) DEFAULT NULL COMMENT '组件',
  `component_name` varchar(20) DEFAULT NULL COMMENT '组件名称',
  `component_type` varchar(20) DEFAULT NULL COMMENT '组件类型',
  `topic` varchar(20) DEFAULT NULL COMMENT 'topic',
  `topic_type` varchar(20) DEFAULT NULL COMMENT 'topic类型',
  `status` varchar(20) DEFAULT 'ON' COMMENT '状态',
  PRIMARY KEY (`id`),
  KEY `component` (`component`,`topic`),
  KEY `idx_topic` (`topic`) USING BTREE
) ENGINE=InnoDB AUTO_INCREMENT=40 DEFAULT CHARSET=utf8

posted @ 2018-08-21 10:47 bronk 阅读(417) 评论(0) 编辑收藏举报

刷新页面返回顶部

bronk

1-监控界面sql保存

1, BufferSize_machine

1), template

2) KafkaSinkNetword

2, BufferSizeTopic

3, Metric_machine

4, Metric_topic

5, QPS_Component

6, QPS_machine

7, Resource_machine

8, alter 报警用, 需要配合设置

9, 需要用到的sql保存

公告