Prometheus连接nVisual实现资产拓扑业务关联分析

你是否想基于机柜视图监控你的设备？
你是否想基于逻辑或物理拓扑监控你的IT设备？
你是否想基于业务流程图监控你的业务？
你是否想同时查看一个设备的位置、拓扑关系、实时状态数据？
……
如果以上有一点是您的需求，通过简单配置你的Prometheus对接nVisual即可快速免费实现。
具体配置步骤如下：
1、在https://cloud.nvisual.com注册免费nVisual云帐号；
2、登录nVisual云台-权限管理-API key-生成并复制API key;
3、在Promeheus的配置文件里增加如下remote wrte配置并重新加载：

remote_write:
  - url: https://cloud.nvisual.com/diagramApi/wapi/v1/prometheus/write
    bearer_token: "你的nVisual API key"
    remote_timeout: 15s
    queue_config:
      capacity: 5000
      max_shards: 1
      min_shards: 1
      max_samples_per_send: 10000
      batch_send_deadline: 15s
      min_backoff: 1s
      max_backoff: 2m
    write_relabel_configs:
      - source_labels: [__name__]
       regex: 'nvisual_device_status|nvisual_port_status' 
       action: keep

4、在Prometheus的rules下创建nvisual_rule.yml配置文件并重新加载：

groups:
  - name: pg_collector_rule
    interval: 10s
    rules:
      - record: nv_pg_collector_rack_remaining
        expr: pg_collector_rack_remaining
  - name: nivsual_device_status
    interval: 10s
    rules:
      - record: nvisual_device_status
        expr: >
          label_replace(up{job=~"snmp|linux|windows|ipmi"},"instance", "$1", "instance", "(.*?)(:\\d+)?$")
          or
          label_replace(probe_success{job=~"blackbox_http|blackbox_icmp"},"instance", "$1", "instance", "(.*?)(:\\d+)?$")

  - name: nvisual_port_status
    interval: 10s
    rules:
      - record: nvisual_port_status
        expr: sum(ifOperStatus) by (ifName, instance)  

5、在alertmanager的alertmanager.yml配置文件增加webhook配置并重新加载：

webhook_configs:
    - url: 'https://cloud.nvisual.com/diagramApi/wapi/v1/webhook/alertmanager' 
      send_resolved: true
      http_config:
        bearer_token: "你的nVisual API key"

6、实现效果截屏