IP	主机名	运行角色
192.168.170.136	hadoop01	namenode datanode resourcemanager nodemanager JournalNode DFSZKFailoverController QuorumPeerMain spark hive
192.168.170.137	hadoop02	namenode datanode resourcemanager nodemanager JournalNode DFSZKFailoverController QuorumPeerMain spark
192.168.170.138	hadoop03	datanode nodemanage JournalNode QuorumPeerMain spark

一、部署 prometheus 环境

1.1 下载安装包

prometheus、alertmanager、pushgateway、node_exporter：https://prometheus.io/download/
prometheus-webhook-dingtalk：https://github.com/timonwong/prometheus-webhook-dingtalk/tree/main
grafana：https://grafana.com/grafana/download
jmx_exporter：https://github.com/prometheus/jmx_exporter

1.2 解压安装

新建一个 /monitor 目录，把上面下载的 tar.gz 包都解压安装在 /monitor 目录下，并重命名如下名字：

1.3 修改配置文件

1.3.1 hadoop-env.sh

修改完后要把这个文件 scp 给各个 Hadoop 节点！

[root@hadoop01 ~]# cd /bigdata/hadoop/server/hadoop-3.2.4/etc/hadoop/
[root@hadoop01 /bigdata/hadoop/server/hadoop-3.2.4/etc/hadoop]# vim hadoop-env.sh
if ! grep -q <<<"$HDFS_NAMENODE_OPTS" jmx_prometheus_javaagent; then
HDFS_NAMENODE_OPTS="$HDFS_NAMENODE_OPTS -javaagent:/monitor/jmx_prometheus_javaagent-0.19.0.jar=30002:/bigdata/hadoop/server/hadoop-3.2.4/prometheus_config.yml"
fi
if ! grep -q <<<"$HDFS_DATANODE_OPTS" jmx_prometheus_javaagent; then
HDFS_DATANODE_OPTS="$HDFS_DATANODE_OPTS -javaagent:/monitor/jmx_prometheus_javaagent-0.19.0.jar=30003:/bigdata/hadoop/server/hadoop-3.2.4/prometheus_config.yml"
fi
if ! grep -q <<<"$YARN_RESOURCEMANAGER_OPTS" jmx_prometheus_javaagent; then
YARN_RESOURCEMANAGER_OPTS="$YARN_RESOURCEMANAGER_OPTS -javaagent:/monitor/jmx_prometheus_javaagent-0.19.0.jar=30004:/bigdata/hadoop/server/hadoop-3.2.4/prometheus_config.yml"
fi
if ! grep -q <<<"$YARN_NODEMANAGER_OPTS" jmx_prometheus_javaagent; then
YARN_NODEMANAGER_OPTS="$YARN_NODEMANAGER_OPTS -javaagent:/monitor/jmx_prometheus_javaagent-0.19.0.jar=30005:/bigdata/hadoop/server/hadoop-3.2.4/prometheus_config.yml"
fi
if ! grep -q <<<"$HDFS_JOURNALNODE_OPTS" jmx_prometheus_javaagent; then
HDFS_JOURNALNODE_OPTS="$HDFS_JOURNALNODE_OPTS -javaagent:/monitor/jmx_prometheus_javaagent-0.19.0.jar=30006:/bigdata/hadoop/server/hadoop-3.2.4/prometheus_config.yml"
fi
if ! grep -q <<<"$HDFS_ZKFC_OPTS" jmx_prometheus_javaagent; then
HDFS_ZKFC_OPTS="$HDFS_ZKFC_OPTS -javaagent:/monitor/jmx_prometheus_javaagent-0.19.0.jar=30007:/bigdata/hadoop/server/hadoop-3.2.4/prometheus_config.yml"
fi
if ! grep -q <<<"$HDFS_HTTPFS_OPTS" jmx_prometheus_javaagent; then
HDFS_HTTPFS_OPTS="$HDFS_HTTPFS_OPTS -javaagent:/monitor/jmx_prometheus_javaagent-0.19.0.jar=30008:/bigdata/hadoop/server/hadoop-3.2.4/prometheus_config.yml"
fi
if ! grep -q <<<"$YARN_PROXYSERVER_OPTS" jmx_prometheus_javaagent; then
YARN_PROXYSERVER_OPTS="$YARN_PROXYSERVER_OPTS -javaagent:/monitor/jmx_prometheus_javaagent-0.19.0.jar=30009:/bigdata/hadoop/server/hadoop-3.2.4/prometheus_config.yml"
fi
if ! grep -q <<<"$MAPRED_HISTORYSERVER_OPTS" jmx_prometheus_javaagent; then
MAPRED_HISTORYSERVER_OPTS="$MAPRED_HISTORYSERVER_OPTS -javaagent:/monitor/jmx_prometheus_javaagent-0.19.0.jar=30010:/bigdata/hadoop/server/hadoop-3.2.4/prometheus_config.yml"
fi

1.3.2 prometheus_config.yml

修改完后要把这个文件 scp 给各个 Hadoop 节点！

[root@hadoop01 ~]# cd /bigdata/hadoop/server/hadoop-3.2.4/
[root@hadoop01 /bigdata/hadoop/server/hadoop-3.2.4]# vim prometheus_config.yml
rules:
- pattern: ".*"

1.3.3 zkServer.sh

修改完后要把这个文件 scp 给各个 zookeeper 节点！

[root@hadoop01 ~]# cd /bigdata/hadoop/zookeeper/zookeeper-3.7.1/bin/
[root@hadoop01 /bigdata/hadoop/zookeeper/zookeeper-3.7.1/bin]# vim zkServer.sh
if [ "x$JMXLOCALONLY" = "x" ]
then
    JMXLOCALONLY=false
fi

JMX_DIR="/monitor"
JVMFLAGS="$JVMFLAGS -javaagent:$JMX_DIR/jmx_prometheus_javaagent-0.19.0.jar=30011:/bigdata/hadoop/zookeeper/zookeeper-3.7.1/prometheus_zookeeper.yaml"

1.3.4 prometheus_zookeeper.yaml

修改完后要把这个文件 scp 给各个 zookeeper 节点！

[root@hadoop01 ~]# cd /bigdata/hadoop/zookeeper/zookeeper-3.7.1/
[root@hadoop01 /bigdata/hadoop/zookeeper/zookeeper-3.7.1]# vim prometheus_zookeeper.yaml 
rules:
  - pattern: "org.apache.ZooKeeperService<name0=ReplicatedServer_id(\\d+)><>(\\w+)"
    name: "zookeeper_$2"
    type: GAUGE
  - pattern: "org.apache.ZooKeeperService<name0=ReplicatedServer_id(\\d+), name1=replica.(\\d+)><>(\\w+)"
    name: "zookeeper_$3"
    type: GAUGE
    labels:
      replicaId: "$2"
  - pattern: "org.apache.ZooKeeperService<name0=ReplicatedServer_id(\\d+), name1=replica.(\\d+), name2=(\\w+)><>(Packets\\w+)"
    name: "zookeeper_$4"
    type: COUNTER
    labels:
      replicaId: "$2"
      memberType: "$3"
  - pattern: "org.apache.ZooKeeperService<name0=ReplicatedServer_id(\\d+), name1=replica.(\\d+), name2=(\\w+)><>(\\w+)"
    name: "zookeeper_$4"
    type: GAUGE
    labels:
      replicaId: "$2"
      memberType: "$3"
  - pattern: "org.apache.ZooKeeperService<name0=ReplicatedServer_id(\\d+), name1=replica.(\\d+), name2=(\\w+), name3=(\\w+)><>(\\w+)"
    name: "zookeeper_$4_$5"
    type: GAUGE
    labels:
      replicaId: "$2"
      memberType: "$3"

  - pattern: "org.apache.ZooKeeperService<name0=StandaloneServer_port(\\d+)><>(\\w+)"
    type: GAUGE
    name: "zookeeper_$2"
  - pattern: "org.apache.ZooKeeperService<name0=StandaloneServer_port(\\d+), name1=InMemoryDataTree><>(\\w+)"
    type: GAUGE
    name: "zookeeper_$2"

1.3.5 alertmanager.yml

[root@hadoop01 ~]# cd /monitor/alertmanager/
[root@hadoop01 /monitor/alertmanager]# ls
alertmanager  alertmanager.yml  amtool  data  LICENSE  NOTICE
[root@hadoop01 /monitor/alertmanager]# vim alertmanager.yml 
global:
  resolve_timeout: 5m

templates:
  - '/monitor/prometheus-webhook-dingtalk/contrib/templates/legacy/*.tmpl'

route:
  group_by: ['job', 'severity']
  group_wait: 30s
  group_interval: 5m
  repeat_interval: 3h
  receiver: 'webhook1'

receivers:
  - name: 'webhook1'
    webhook_configs:
    - url: 'http://192.168.170.136:8060/dingtalk/webhook1/send'
      send_resolved: true

1.3.6 prometheus.yml

[root@hadoop01 ~]# cd /monitor/prometheus
[root@hadoop01 /monitor/prometheus]# ls
console_libraries  consoles  data  LICENSE  NOTICE  prometheus  prometheus.yml  promtool  rule
[root@hadoop01 /monitor/prometheus]# vim prometheus.yml 
# my global config
global:
  scrape_interval: 30s # Set the scrape interval to every 15 seconds. Default is every 1 minute.
  evaluation_interval: 30s # Evaluate rules every 15 seconds. The default is every 1 minute.

# Alertmanager configuration
alerting:
  alertmanagers:
    - static_configs:
        - targets: ['192.168.170.136:9093']

# Load rules once and periodically evaluate them according to the global 'evaluation_interval'.
rule_files:
  - "rule/*.yml"

# A scrape configuration containing exactly one endpoint to scrape:
# Here it's Prometheus itself.
scrape_configs:
  - job_name: "prometheus"
    scrape_interval: 30s
    static_configs:
      - targets: ["hadoop01:9090"]

  # zookeeper 集群配置
  - job_name: "zookeeper"
    scrape_interval: 30s
    static_configs:
      - targets: ['hadoop01:30011', 'hadoop02:30011', 'hadoop03:30011']

  # node_exporter 配置
  - job_name: "pushgatewawy"
    scrape_interval: 30s
    static_configs:
      - targets: ["hadoop01:9091"]
                                                                        
  # node_exporter 配置
  - job_name: "node_exporter"
    scrape_interval: 30s
    static_configs:
      - targets: ['hadoop01:9100', 'hadoop02:9100', 'hadoop03:9100']

  - job_name: " namenode "
    scrape_interval: 30s
    static_configs:
      - targets: ['hadoop01:30002', 'hadoop02:30002']
#        labels:
#          instance: namenode 服务器

  - job_name: "datanode"
    scrape_interval: 30s
    static_configs:
      - targets: ['hadoop01:30003', 'hadoop02:30003', 'hadoop03:30003']

  - job_name: "resourcemanager"
    scrape_interval: 30s
    static_configs:
      - targets: ['hadoop01:30004', 'hadoop02:30004']

  - job_name: "nodemanager"
    scrape_interval: 30s
    static_configs:
      - targets: ['hadoop01:30005', 'hadoop02:30005', 'hadoop03:30005']

  - job_name: "journalnode"
    scrape_interval: 30s
    static_configs:
      - targets: ['hadoop01:30006', 'hadoop02:30006', 'hadoop03:30006']

  - job_name: "zkfc"
    scrape_interval: 30s
    static_configs:
      - targets: ['hadoop01:30007', 'hadoop02:30007']

  - job_name: "jobhistoryserver"
    scrape_interval: 30s
    static_configs:
      - targets: ["hadoop01:30010"]
      
  - job_name: "spark_master"
    scrape_interval: 30s
    static_configs:
      - targets: ['hadoop01:30012', 'hadoop02:30012']

  - job_name: "spark_worker"
    scrape_interval: 30s
    static_configs:
      - targets: ['hadoop01:30013', 'hadoop02:30013', 'hadoop03:30013'] 
  
  - job_name: "hive_metastore"
    scrape_interval: 30s
    static_configs:
      - targets: ["hadoop01:30014"]

  - job_name: "hive_hs2"
    scrape_interval: 30s
    static_configs:
      - targets: ["hadoop01:30015"]

1.3.7 config.yml

[root@hadoop01 ~]# cd /monitor/prometheus-webhook-dingtalk/
[root@hadoop01 /monitor/prometheus-webhook-dingtalk]# ls
config.example.yml  config.yml  contrib  LICENSE  nohup.out  prometheus-webhook-dingtalk
[root@hadoop01 /monitor/prometheus-webhook-dingtalk]# vim config.yml
## Request timeout
# timeout: 5s

## Uncomment following line in order to write template from scratch (be careful!)
#no_builtin_template: true

## Customizable templates path
templates:
  - /monitor/prometheus-webhook-dingtalk/contrib/templates/legacy/template.tmpl

## You can also override default template using `default_message`
## The following example to use the 'legacy' template from v0.3.0
#default_message:
#  title: '{{ template "legacy.title" . }}'
#  text: '{{ template "legacy.content" . }}'

## Targets, previously was known as "profiles"
targets:
  webhook1:
    url: https://oapi.dingtalk.com/robot/send?access_token=0d6c5dc25fa3f79cf2f83c92705fe4594dcxxx
    # secret for signature
    secret: SECecdbfff858ab8f3195dc34b7e225fee9341bc9xxx
    message:
      title: '{{ template "ops.title" . }}'
      text: '{{ template "ops.content" . }}'

1.3.8 template.tmpl

[root@hadoop01 ~]# cd /monitor/prometheus-webhook-dingtalk/contrib/templates/legacy/
[root@hadoop01 /monitor/prometheus-webhook-dingtalk/contrib/templates/legacy]# vim template.tmpl
{{ define "__subject" }}[{{ .Status | toUpper }}{{ if eq .Status "firing" }}:{{ .Alerts.Firing | len }}{{ end }}]
{{ end }}

{{ define "__alert_list" }}{{ range . }}
---
    **告警类型**: {{ .Labels.alertname }}
    **告警级别**: {{ .Labels.severity }}
    **故障主机**: {{ .Labels.instance }}
    **告警信息**: {{ .Annotations.description }}
    **触发时间**: {{ (.StartsAt.Add 28800e9).Format "2006-01-02 15:04:05" }}
{{ end }}{{ end }}

{{ define "__resolved_list" }}{{ range . }}
---
    **告警类型**: {{ .Labels.alertname }}
    **告警级别**: {{ .Labels.severity }}
    **故障主机**: {{ .Labels.instance }}
    **触发时间**: {{ (.StartsAt.Add 28800e9).Format "2006-01-02 15:04:05" }}
    **恢复时间**: {{ (.EndsAt.Add 28800e9).Format "2006-01-02 15:04:05" }}
{{ end }}{{ end }}


{{ define "ops.title" }}
{{ template "__subject" . }}
{{ end }}

{{ define "ops.content" }}
{{ if gt (len .Alerts.Firing) 0 }}
**====侦测到{{ .Alerts.Firing | len  }}个故障====**
{{ template "__alert_list" .Alerts.Firing }}
---
{{ end }}

{{ if gt (len .Alerts.Resolved) 0 }}
**====恢复{{ .Alerts.Resolved | len  }}个故障====**
{{ template "__resolved_list" .Alerts.Resolved }}
{{ end }}
{{ end }}

{{ define "ops.link.title" }}{{ template "ops.title" . }}{{ end }}
{{ define "ops.link.content" }}{{ template "ops.content" . }}{{ end }}
{{ template "ops.title" . }}
{{ template "ops.content" . }}

1.3.9 告警规则

在第二点下面的文件里下载即可。

1.3.10 /etc/profile

修改完后要把这个文件 scp 给各个 Hadoop 节点！

# JDK 1.8
JAVA_HOME=/usr/java/jdk1.8.0_381
PATH=$PATH:$JAVA_HOME/bin
CLASSPATH=.:$JAVA_HOME/lib
export JAVA_HOME PATH CLASSPATH

# hadoop
export HADOOP_HOME=/bigdata/hadoop/server/hadoop-3.2.4/
export PATH=$PATH:$HADOOP_HOME/bin:$HADOOP_HOME/sbin

# spark
export SPARK_HOME=/bigdata/spark-3.2.4
export HADOOP_CONF_DIR=$HADOOP_HOME/etc/hadoop
export PYSPARK_PYTHON=/usr/local/anaconda3/envs/pyspark/bin/python3.10
export LD_LIBRARY_PATH=$LD_LIBRARY_PATH:$HADOOP_HOME/lib/native

1.3.11 prometheus_spark.yml

修改完后要把这个文件 scp 给各个 spark 节点！

(base) [root@hadoop01 ~]# cd /bigdata/spark-3.2.4/
(base) [root@hadoop01 /bigdata/spark-3.2.4]# vim prometheus_spark.yml
rules:

  # These come from the master
  # Example: master.aliveWorkers
  - pattern: "metrics<name=master\\.(.*), type=counters><>Value"
    name: spark_master_$1

  # These come from the worker
  # Example: worker.coresFree
  - pattern: "metrics<name=worker\\.(.*), type=counters><>Value"
    name: spark_worker_$1

  # These come from the application driver
  # Example: app-20160809000059-0000.driver.DAGScheduler.stage.failedStages
  - pattern: "metrics<name=(.*)\\.driver\\.(DAGScheduler|BlockManager|jvm)\\.(.*), type=gauges><>Value"
    name: spark_driver_$2_$3
    type: GAUGE
    labels:
      app_id: "$1"

  # These come from the application driver
  # Emulate timers for DAGScheduler like messagePRocessingTime
  - pattern: "metrics<name=(.*)\\.driver\\.DAGScheduler\\.(.*), type=counters><>Count"
    name: spark_driver_DAGScheduler_$2_total
    type: COUNTER
    labels:
      app_id: "$1"

  - pattern: "metrics<name=(.*)\\.driver\\.HiveExternalCatalog\\.(.*), type=counters><>Count"
    name: spark_driver_HiveExternalCatalog_$2_total
    type: COUNTER
    labels:
      app_id: "$1"

  # These come from the application driver
  # Emulate histograms for CodeGenerator
  - pattern: "metrics<name=(.*)\\.driver\\.CodeGenerator\\.(.*), type=counters><>Count"
    name: spark_driver_CodeGenerator_$2_total
    type: COUNTER
    labels:
      app_id: "$1"

  # These come from the application driver
  # Emulate timer (keep only count attribute) plus counters for LiveListenerBus
  - pattern: "metrics<name=(.*)\\.driver\\.LiveListenerBus\\.(.*), type=counters><>Count"
    name: spark_driver_LiveListenerBus_$2_total
    type: COUNTER
    labels:
      app_id: "$1"

  # Get Gauge type metrics for LiveListenerBus
  - pattern: "metrics<name=(.*)\\.driver\\.LiveListenerBus\\.(.*), type=gauges><>Value"
    name: spark_driver_LiveListenerBus_$2
    type: GAUGE
    labels:
      app_id: "$1"

  # These come from the application driver if it's a streaming application
  # Example: app-20160809000059-0000.driver.com.example.ClassName.StreamingMetrics.streaming.lastCompletedBatch_schedulingDelay
  - pattern: "metrics<name=(.*)\\.driver\\.(.*)\\.StreamingMetrics\\.streaming\\.(.*), type=gauges><>Value"
    name: spark_driver_streaming_$3
    labels:
      app_id: "$1"
      app_name: "$2"

  # These come from the application driver if it's a structured streaming application
  # Example: app-20160809000059-0000.driver.spark.streaming.QueryName.inputRate-total
  - pattern: "metrics<name=(.*)\\.driver\\.spark\\.streaming\\.(.*)\\.(.*), type=gauges><>Value"
    name: spark_driver_structured_streaming_$3
    labels:
      app_id: "$1"
      query_name: "$2"

  # These come from the application executors
  # Examples:
  #  app-20160809000059-0000.0.executor.threadpool.activeTasks (value)
  #  app-20160809000059-0000.0.executor.JvmGCtime (counter)

  # filesystem metrics are declared as gauge metrics, but are actually counters
  - pattern: "metrics<name=(.*)\\.(.*)\\.executor\\.filesystem\\.(.*), type=gauges><>Value"
    name: spark_executor_filesystem_$3_total
    type: COUNTER
    labels:
      app_id: "$1"
      executor_id: "$2"

  - pattern: "metrics<name=(.*)\\.(.*)\\.executor\\.(.*), type=gauges><>Value"
    name: spark_executor_$3
    type: GAUGE
    labels:
      app_id: "$1"
      executor_id: "$2"

  - pattern: "metrics<name=(.*)\\.(.*)\\.executor\\.(.*), type=counters><>Count"
    name: spark_executor_$3_total
    type: COUNTER
    labels:
      app_id: "$1"
      executor_id: "$2"

  - pattern: "metrics<name=(.*)\\.(.*)\\.ExecutorMetrics\\.(.*), type=gauges><>Value"
    name: spark_executor_$3
    type: GAUGE
    labels:
      app_id: "$1"
      executor_id: "$2"

  # These come from the application executors
  # Example: app-20160809000059-0000.0.jvm.threadpool.activeTasks
  - pattern: "metrics<name=(.*)\\.([0-9]+)\\.(jvm|NettyBlockTransfer)\\.(.*), type=gauges><>Value"
    name: spark_executor_$3_$4
    type: GAUGE
    labels:
      app_id: "$1"
      executor_id: "$2"

  - pattern: "metrics<name=(.*)\\.([0-9]+)\\.HiveExternalCatalog\\.(.*), type=counters><>Count"
    name: spark_executor_HiveExternalCatalog_$3_total
    type: COUNTER
    labels:
      app_id: "$1"
      executor_id: "$2"

  # These come from the application driver
  # Emulate histograms for CodeGenerator
  - pattern: "metrics<name=(.*)\\.([0-9]+)\\.CodeGenerator\\.(.*), type=counters><>Count"
    name: spark_executor_CodeGenerator_$3_total
    type: COUNTER
    labels:
      app_id: "$1"
      executor_id: "$2"

1.3.12 spark-env.sh

修改完后要把这个文件 scp 给各个 spark 节点！

(base) [root@hadoop01 ~]# cd /bigdata/spark-3.2.4/conf/
(base) [root@hadoop01 /bigdata/spark-3.2.4/conf]# vim spark-env.sh
export SPARK_MASTER_OPTS="$SPARK_MASTER_OPTS -javaagent:/monitor/jmx_prometheus_javaagent-0.19.0.jar=30012:/bigdata/spark-3.2.4/prometheus_spark.yml"
export SPARK_WORKER_OPTS="$SPARK_WORKER_OPTS -javaagent:/monitor/jmx_prometheus_javaagent-0.19.0.jar=30013:/bigdata/spark-3.2.4/prometheus_spark.yml"

1.3.13 hive

(base) [root@hadoop01 ~]# cd /bigdata/apache-hive-3.1.2/
(base) [root@hadoop01 /bigdata/apache-hive-3.1.2]# vim bin/hive
···
if [[ "$SERVICE" =~ ^(help|version|orcfiledump|rcfilecat|schemaTool|cleardanglingscratchdir|metastore|beeline|llapstatus|llap)$ ]] ; then
  # 如果是 metastore 服务，则修改 HADOOP_CLIENT_OPTS
  if [[ "$SERVICE" == "metastore" ]] ; then
    export HADOOP_CLIENT_OPTS="$HADOOP_CLIENT_OPTS -javaagent:/monitor/jmx_prometheus_javaagent-0.19.0.jar=30014:/bigdata/apache-hive-3.1.2/prometheus_metastore.yaml"
  fi
  SKIP_HBASECP=true
fi

···
if [[ "$SERVICE" =~ ^(hiveserver2|beeline|cli)$ ]] ; then
  # 如果是 hiveserver2 服务，则修改 HADOOP_CLIENT_OPTS
  if [[ "$SERVICE" == "hiveserver2" ]] ; then
    export HADOOP_CLIENT_OPTS="$HADOOP_CLIENT_OPTS -javaagent:/monitor/jmx_prometheus_javaagent-0.19.0.jar=30015:/bigdata/apache-hive-3.1.2/prometheus_hs2.yaml"
  fi
  # If process is backgrounded, don't change terminal settings
  if [[ ( ! $(ps -o stat= -p $$) =~ "+" ) && ! ( -p /dev/stdin ) && ( ! $(ps -o tty= -p $$) =~ "?" ) ]]; then
    export HADOOP_CLIENT_OPTS="$HADOOP_CLIENT_OPTS -Djline.terminal=jline.UnsupportedTerminal"
  fi
fi
···

1.3.14 prometheus_metastore.yaml、prometheus_hs2.yaml

(base) [root@hadoop01 /bigdata/apache-hive-3.1.2]# vim prometheus_metastore.yaml 
---
startDelaySeconds: 0
ssl: false
lowercaseOutputName: false
lowercaseOutputLabelNames: false
rules:
- pattern: ".*"

(base) [root@hadoop01 /bigdata/apache-hive-3.1.2]# vim prometheus_hs2.yaml 
---
startDelaySeconds: 0
ssl: false
lowercaseOutputName: false
lowercaseOutputLabelNames: false
rules:
- pattern: ".*"

1.4 创建 systemd 服务

1.4.1 创建 prometheus 用户

各个节点都需要创建！！！（也可以不创建 prometheus 用户，把后面 service 文件的 prometheus 改为 root 即可！）

useradd -M -s /usr/sbin/nologin prometheus
chown -R prometheus:prometheus /monitor

1.4.2 alertmanager.service

[root@hadoop01 ~]# vim /usr/lib/systemd/system/alertmanager.service 
[Unit]
Description=Alertmanager
Documentation=https://prometheus.io/docs/alerting/alertmanager/
After=network-online.target
Wants=network-online.target

[Service]
User=prometheus
Group=prometheus
Type=simple
ExecStart=/monitor/alertmanager/alertmanager \
    --config.file=/monitor/alertmanager/alertmanager.yml \
    --storage.path=/monitor/alertmanager/data \
    --web.listen-address=0.0.0.0:9093
ExecReload=/bin/kill -HUP $MAINPID
Restart=always

[Install]
WantedBy=multi-user.target

1.4.3 prometheus.service

[root@hadoop01 ~]# vim /usr/lib/systemd/system/prometheus.service 
[Unit]
Description=Prometheus Server
Documentation=https://prometheus.io/docs/introduction/overview/
After=network-online.target

[Service]
Type=simple
User=prometheus
Group=prometheus
WorkingDirectory=/monitor/prometheus
ExecStart=/monitor/prometheus/prometheus \
  --web.listen-address=0.0.0.0:9090 \
  --storage.tsdb.path=/monitor/prometheus/data \
  --storage.tsdb.retention.time=30d \
  --config.file=prometheus.yml \
  --web.enable-lifecycle
ExecReload=/bin/kill -s HUP $MAINPID
ExecStop=/bin/kill -s QUIT $MAINPID
Restart=on-failure

[Install]
WantedBy=multi-user.target

1.4.4 node_exporter.service

[root@hadoop01 ~]# vim /usr/lib/systemd/system/node_exporter.service 
[Unit]
Description=Node Exporter
Documentation=https://github.com/prometheus/node_exporter
After=network-online.target
Wants=network-online.target

[Service]
User=prometheus
Group=prometheus
Type=simple
ExecStart=/monitor/node_exporter/node_exporter

[Install]
WantedBy=multi-user.target

1.4.5 pushgateway.service

[root@hadoop01 ~]# vim /usr/lib/systemd/system/pushgateway.service 
[Unit]
Description=Pushgateway Server
Documentation=https://github.com/prometheus/pushgateway
After=network-online.target
Wants=network-online.target

[Service]
User=prometheus
Group=prometheus
Type=simple
ExecStart=/monitor/pushgateway/pushgateway \
          --web.listen-address=:9091 \
          --web.telemetry-path=/metrics
Restart=always

[Install]
WantedBy=multi-user.target

1.4.6 grafana.service

[root@hadoop01 ~]# vim /usr/lib/systemd/system/grafana.service 
[Unit]
Description=Grafana Server
Documentation=http://docs.grafana.org
After=network-online.target
Wants=network-online.target

[Service]
Type=simple
User=prometheus
Group=prometheus
ExecStart=/monitor/grafana/bin/grafana-server \
    --config=/monitor/grafana/conf/defaults.ini \
    --homepath=/monitor/grafana
Restart=on-failure
RestartSec=10
StandardOutput=syslog
StandardError=syslog
SyslogIdentifier=grafana
Environment=GRAFANA_HOME=/monitor/grafana \
            GRAFANA_USER=prometheus \
            GRAFANA_GROUP=prometheus

[Install]
WantedBy=multi-user.target

1.5 启动服务

把上述服务启动即可！

注意：prometheus-webhook-dingtalk 服务需要用下面方式启动：

cd /monitor/prometheus-webhook-dingtalk/
nohup ./prometheus-webhook-dingtalk --web.listen-address="0.0.0.0:8060" --config.file="/monitor/prometheus-webhook-dingtalk/config.yml" &

二、补充

2.1 告警规则和 grafana 仪表盘文件下载

[root@hadoop01 ~]# cd /monitor/prometheus/
[root@hadoop01 /monitor/prometheus]# ls
console_libraries  consoles  data  LICENSE  NOTICE  prometheus  prometheus.yml  promtool  rule
[root@hadoop01 /monitor/prometheus]# ls rule/
HDFS.yml  node.yml  spark_master.yml  spark_worker.yml  yarn.yml  zookeeper.yml

文件下载链接：【免费】prometheus告警规则文件和grafana仪表盘文件资源-CSDN文库

2.2 服务进程监控脚本

2.2.1 prometheus-webhook-dingtalk

[root@hadoop01 ~]# cd /monitor/prometheus-webhook-dingtalk/
[root@hadoop01 /monitor/prometheus-webhook-dingtalk]# vim monitor_prometheus_webhook_dingtalk.sh
#!/bin/bash

# 获取当前系统时间
current_time=$(date "+%Y-%m-%d %H:%M:%S")

# 定义日志文件路径
log_file="/monitor/prometheus-webhook-dingtalk/monitor.log"

echo "[$current_time] Checking if prometheus-webhook-dingtalk process is running..." >> $log_file

# 检查进程是否在运行
if ! /usr/bin/pgrep -fx "/monitor/prometheus-webhook-dingtalk/prometheus-webhook-dingtalk --web.listen-address=0.0.0.0:8060 --config.file=/monitor/prometheus-webhook-dingtalk/config.yml" >> $log_file; then
    echo "[$current_time] prometheus-webhook-dingtalk process is not running. Starting it now..." >> $log_file
    # 使用绝对路径和 nohup 在后台启动进程
    /usr/bin/nohup /monitor/prometheus-webhook-dingtalk/prometheus-webhook-dingtalk --web.listen-address="0.0.0.0:8060" --config.file="/monitor/prometheus-webhook-dingtalk/config.yml" >> /monitor/prometheus-webhook-dingtalk/output.log 2>&1 &

else
    echo "[$current_time] prometheus-webhook-dingtalk process is running." >> $log_file
fi

[root@hadoop01 /monitor/prometheus-webhook-dingtalk]# chmod 777 monitor_prometheus_webhook_dingtalk.sh 

[root@hadoop01 /monitor/prometheus-webhook-dingtalk]# crontab -e
* * * * * /usr/bin/bash /monitor/prometheus-webhook-dingtalk/monitor_prometheus_webhook_dingtalk.sh

2.2.2 hive

(base) [root@hadoop01 ~]# cd /bigdata/apache-hive-3.1.2/
(base) [root@hadoop01 /bigdata/apache-hive-3.1.2]# vim monitor_hive.sh
#!/bin/bash

# 获取当前系统时间
current_time=$(date "+%Y-%m-%d %H:%M:%S")

# 定义日志文件路径
log_file_metastore="/bigdata/apache-hive-3.1.2/monitor_metastore.log"
log_file_hs2="/bigdata/apache-hive-3.1.2/monitor_hs2.log"

echo "[$current_time] Checking if hive metastore and hs2 processes are running..." 

# 检查 Hive Metastore 是否在运行
echo "[$current_time] Checking if hive metastore process is running..." >> $log_file_metastore
if ! /usr/bin/pgrep -f "hive-metastore-3.1.2.jar" >> $log_file_metastore; then
    echo "[$current_time] hive metastore process is not running. Starting it now..." >> $log_file_metastore
    # 使用绝对路径和 nohup 在后台启动进程
    /usr/bin/nohup /bigdata/apache-hive-3.1.2/bin/hive --service metastore >> /bigdata/apache-hive-3.1.2/metastore_output.log 2>&1 &
    # 等待一点时间以确保 metastore 完全启动
    sleep 30
else
    echo "[$current_time] hive metastore process is running." >> $log_file_metastore
fi

# 检查 HiveServer2 是否在运行
echo "[$current_time] Checking if hive hs2 process is running..." >> $log_file_hs2
if ! /usr/bin/pgrep -f "HiveServer2" >> $log_file_hs2; then
    echo "[$current_time] hive hs2 process is not running. Starting it now..." >> $log_file_hs2
    # 使用绝对路径和 nohup 在后台启动进程
    /usr/bin/nohup /bigdata/apache-hive-3.1.2/bin/hive --service hiveserver2 >> /bigdata/apache-hive-3.1.2/hs2_output.log 2>&1 &
else
    echo "[$current_time] hive hs2 process is running." >> $log_file_hs2
fi

(base) [root@hadoop01 /bigdata/apache-hive-3.1.2]# chmod 777 montior_metastore.sh 

(base) [root@hadoop01 /bigdata/apache-hive-3.1.2]# crontab -e
* * * * * /usr/bin/bash /bigdata/apache-hive-3.1.2/monitor_hive.sh

2.2.3 日志切割

[root@hadoop01 ~]# vim /etc/logrotate.d/prometheus-webhook-dingtalk
/monitor/prometheus-webhook-dingtalk/monitor.log \
/bigdata/apache-hive-3.1.2/monitor_metastore.log \
/bigdata/apache-hive-3.1.2/monitor_hs2.log {
    daily
    rotate 7
    size 150M
    compress
    maxage 30
    missingok
    notifempty
    create 0644 root root
    copytruncate
}

# 测试调式 logrotate 配置
[root@hadoop01 ~]# logrotate -d /etc/logrotate.d/prometheus-webhook-dingtalk
# 手动执行日志轮换
logrotate -f /etc/logrotate.d/prometheus-webhook-dingtalk

2.3 常用命令

# 检查 prometheus 配置文件，包括告警规则文件
[root@hadoop01 ~]# cd /monitor/prometheus
./promtool check config prometheus.yml

# 重启 prometheus 配置
curl -X POST http://localhost:9090/-/reload

# 测试发送信息到机器人
curl 'https://oapi.dingtalk.com/robot/send?access_token=xxxxxxxx' \
 -H 'Content-Type: application/json' \
 -d '{"msgtype": "text","text": {"content":"我就是我, 是不一样的烟火"}}'

2.4 参考文档

Hadoop 官方监控指标：https://hadoop.apache.org/docs/r3.2.4/hadoop-project-dist/hadoop-common/Metrics.html
阿里云监控指标：https://help.aliyun.com/zh/emr/emr-on-ecs/user-guide/hdfs-metrics?spm=a2c4g.11186623.0.0.11ba6daalnmBWn
阿里云 grafana 仪表盘：https://help.aliyun.com/document_detail/2326798.html?spm=a2c4g.462292.0.0.4c4c5d35uXCP6k#section-1bn-bzq-fw3
jmx_exporter 配置文件参考：https://github.com/prometheus/jmx_exporter/tree/main/example_configs
钉钉机器人文档：https://open.dingtalk.com/document/robots/custom-robot-access

2.5 内网环境

2.5.1 nginx（公网）

需要一台可以访问公网的 nginx 服务器来代理钉钉 api：

[root@idc-master-02 ~]# cat /etc/nginx/conf.d/uat-prometheus-webhook-dingtalk.conf
server {
    listen 30080;

    location /robot/send {
        proxy_pass https://oapi.dingtalk.com;
        proxy_set_header Host oapi.dingtalk.com;
        proxy_set_header X-Real-IP $remote_addr;
        proxy_set_header X-Forwarded-For $proxy_add_x_forwarded_for;
        proxy_set_header X-Forwarded-Proto $scheme;
    }
}

2.5.2 config

[root@localhost-13 ~]# cat /opt/prometheus-webhook-dingtalk/config.yml 

targets:
  webhook1:
    url: http://10.0.4.11:30080/robot/send?access_token=0d6c5dc25fa3f79cf2f83c92705fe4594dcc5b3xxx
    secret: SECecdbfff858ab8f3195dc34b7e225fee93xxx
    message:
      title: '{{ template "ops.title" . }}'
      text: '{{ template "ops.content" . }}'

2.5.3 /etc/hosts

[root@localhost-13 ~]# cat /etc/hosts
127.0.0.1   localhost localhost.localdomain localhost4 localhost4.localdomain4
::1         localhost localhost.localdomain localhost6 localhost6.localdomain6
10.0.4.11 oapi.dingtalk.com