通过VK获取实例监控指标

1# 指标的帮助说明
2# 指标名称 指标类型
3指标名称{标签，标签，...} 指标取值 指标时间戳
4指标名称{标签，标签，...} 指标取值 指标时间戳
5
6# 指标的帮助说明
7# 指标名称 指标类型
8指标名称{标签，标签，...} 指标取值 指标时间戳
9指标名称{标签，标签，...} 指标取值 指标时间戳

1job_name: 'cadvisor-gpu'
2scheme: https
3tls_config:
4  ca_file: /var/run/secrets/kubernetes.io/serviceaccount/ca.crt
5  insecure_skip_verify: true
6bearer_token_file: /var/run/secrets/kubernetes.io/serviceaccount/token
7metrics_path: /metrics/cadvisor
8kubernetes_sd_configs:
9  - role: node
10relabel_configs:
11  - action: labelmap
12    regex: __meta_kubernetes_node_label_(.+)
13  - source_labels: [ instance ]
14    action: replace
15    target_label: node
16metric_relabel_configs:
17  - source_labels: [__name__]
18    action: keep
19    regex: (machine_(cpu_cores|memory_bytes)|container_(cpu_usage_seconds_total|fs_limit_bytes|fs_reads_bytes_total|
20       fs_usage_bytes|fs_writes_bytes_total|memory_working_set_bytes|network_receive_bytes_total|
21       network_receive_packets_dropped_total|network_receive_packets_total|network_transmit_bytes_total|
22       network_transmit_packets_dropped_total|network_transmit_packets_total)|DCGM_(FI_DEV_GPU_UTIL|FI_DEV_GPU_TEMP|FI_DEV_FB_USED|FI_DEV_FB_FREE|FI_DEV_POWER_USAGE|FI_PROF_SM_ACTIVE|FI_PROF_PCIE_RX_BYTES|FI_PROF_PCIE_TX_BYTES))

1Host:  {KubeletServer} 
2Port:  10250
3URI:   GET /metrics/cadvisor
4Param：无

1./kubectl --kubeconfig={kubeconfig_path} get --raw "/api/v1/nodes/{vk_bci_kubelet_name}/proxy/metrics/cadvisor"
2// {vk_bci_kubelet_name} vk 的 nodeName

1global:
2  scrape_interval: 15s # Set the scrape interval to every 15 seconds. Default is every 1 minute.
3  evaluation_interval: 15s # Evaluate rules every 15 seconds. The default is every 1 minute.
4  # scrape_timeout is set to the global default (10s).
5
6# Alertmanager configuration
7alerting:
8  alertmanagers:
9    - static_configs:
10        - targets:
11          # - alertmanager:9093
12
13# Load rules once and periodically evaluate them according to the global 'evaluation_interval'.
14rule_files:
15  # - "first_rules.yml"
16  # - "second_rules.yml"
17
18# A scrape configuration containing exactly one endpoint to scrape:
19# Here it's Prometheus itself.
20scrape_configs:
21  # The job name is added as a label `job=<job_name>` to any timeseries scraped from this config.
22  - job_name: "prometheus"
23    static_configs:
24      - targets: ["localhost:9090"]
25  - job_name: "cadvisor"
26    static_configs:
27    - targets: ['47.92.228.38:10250']  // 此处 47.92.228.38 是 vk 因特网可访问的 ip，10250 是 vk 的 https server 端口（注意不是 http）
28    scheme: https
29    metrics_path: /metrics/cadvisor
30    tls_config:
31      insecure_skip_verify: true    // 作为一个demo示例，关闭 tls 协议
32    authorization:
33      credentials_file: /home/work/prometheus/token  // 可以正常访问 node server 的 bearer token
34    relabel_configs:
35      - action: labelmap
36        regex: __meta_kubernetes_node_label_(.+)

1./prometheus --config.file="prometheus.yml"