人工智能AI:Keras PyTorch MXNet TensorFlow PaddlePaddle 深度学习实战(不定时更新)
1.node_exporter下载: https://prometheus.io/download/
2.修改prometheus配置文件 job_name: 'prometheus':prometheus配置 job_name: 'pushgateway':pushgateway配置 job_name: 'node1':node_exporter在master节点的配置 job_name: 'node2':node_exporter在slave1节点的配置 job_name: 'node3':node_exporter在slave2节点的配置 vim /home/hadoop/prometheus/prometheus.yml # A scrape configuration containing exactly one endpoint to scrape: # Here it's Prometheus itself. scrape_configs: # The job name is added as a label `job=<job_name>` to any timeseries scraped from this config. - job_name: 'prometheus' # metrics_path defaults to '/metrics' # scheme defaults to 'http'. static_configs: - targets: ['xxx.xxx.xx.xx:9090'] labels: instance: 'prometheus' - job_name: 'pushgateway' static_configs: - targets: ['xxx.xxx.xx.xx:9091'] labels: instance: 'pushgateway' - job_name: 'slave2' static_configs: - targets: ['xxx.xxx.xx.xx:9100'] labels: instance: 'slave2' - job_name: 'slave1' static_configs: - targets: ['xxx.xxx.xx.xx:9100'] labels: instance: 'slave1' - job_name: 'master' static_configs: - targets: ['xxx.xxx.xx.xx:9100'] labels: instance: 'master' 3.多台机器 每台机器都 安装和启动 node_exporter tar zxvf node_exporter-1.0.1.linux-amd64.tar.gz mv node_exporter-1.0.1.linux-amd64 node_exporter cd /home/hadoop/node_exporter ./node_exporter &(用了&之后即启动了之后再按一次回车就可以退出打印台) ps aux|grep node_exporter netstat -anp |grep 9100 http://每台机器的IP:9100(查看UI界面)
4.启动 prometheus cd /home/hadoop/prometheus ./prometheus --config.file=prometheus.yml &(用了&之后即启动了之后再按一次回车就可以退出打印台) http://xxx.xxx.xx.xx:9090 打开UI界面 ps aux|grep prometheus netstat -anp |grep 9090
5.下载 dashboards 仪表盘:Node Exporter for Prometheus Dashboard CN
6.安装 grafana tar zxvf grafana-7.1.0.linux-amd64.tar.gz mv grafana-7.1.0 grafana cd /home/hadoop/grafana 启动 grafana cd /home/hadoop/grafana/bin ./grafana-server &(用了&之后即启动了之后再按一次回车就可以退出打印台) ./grafana-server -config="/home/hadoop/grafana/conf/defaults.ini" ps aux|grep grafana-server netstat -anp |grep 3000 登录 grafana http://xxx.xxx.xx.xx:3000 使用默认用户名密码(admin/admin) 登录报错1: Unauthorized Invalid or expired reset password code path=/api/user/password/reset status=400 path=/api/user/password status=401 登录报错2: grafana登录页面 输入账号密码后又回到登录界面 解决: 设置免密码免用户登录,并且是Admin权限登录 vim /home/hadoop/grafana/conf/defaults.ini [auth.anonymous] #启用匿名访问 enabled = true #默认false #为未经身份验证的用户指定角色 org_role = Admin #默认Viewer