一、部署服务发现
1.1 基于文件的服务发现
基于文件的服务发现是仅仅略优于静态配置的服务发现方式,它不依赖于任何平台或第三方服务,因而也是最为简单和通用的实现方式。
Prometheus Server 会定期从文件中加载 Target 信息,文件可使用 YAML 和 JSON 格式,它含有定义的 Target 列表,以及可选的标签信息。
(1)创建用于服务发现的文件,在文件中配置所需的 target
cd /usr/local/prometheus
mkdir targets
vim targets/node-exporter.yaml
- targets:
- 192.168.80.130:9100
labels:
from: node-exporter
vim targets/mysqld-exporter.yaml
- targets:
- 192.168.80.160:9104
labels:
from: mysqld-exporter
(2)修改 prometheus 配置文件,发现 target 的配置,定义在配置文件的 job 之中
vim /usr/local/prometheus/prometheus.yml
......
scrape_configs:
- job_name: nodes
file_sd_configs: #指定使用文件服务发现
- files: #指定要加载的文件列表
- targets/node*.yaml #文件加载支持通配符
refresh_interval: 2m #每隔 2 分钟重新加载一次文件中定义的 Targets,默认为 5m
- job_name: mysqld
file_sd_configs:
- files:
- targets/mysqld*.yaml
refresh_interval: 2m
systemctl reload prometheus
浏览器查看 Prometheus 页面的 Status -> Targets
1.2 基于 Consul 的服务发现
Consul 是一款基于 golang 开发的开源工具,主要面向分布式,服务化的系统提供服务注册、服务发现和配置管理的功能。
提供服务注册/发现、健康检查、Key/Value存储、多数据中心和分布式一致性保证等功能。下载地址:https://www.consul.io/downloads/
(1)部署 Consul 服务
cd /opt/
mkdir cnsul
unzip consul_1.9.2_linux_amd64.zip
mv consul /usr/local/bin/
#创建 Consul 服务的数据目录和配置目录
mkdir data logs conf
#使用 server 模式启动 Consul 服务
consul agent -server -bootstrap -ui -data-dir=./data -config-dir=./conf -bind=192.168.136.160 -client=0.0.0.0 -node=consul-server01 &> ./logs/consul.log &
#查看 consul 集群成员
consul members
(2)在 Consul 上注册 Services
#在配置目录中添加文件
cd conf/
vim nodes.json
{
"services": [
{
"id": "node_exporter-node01",
"name": "node01",
"address": "192.168.136.130",
"port": 9100,
"tags": ["nodes"],
"checks": [{
"http": "http://192.168.136.130:9100/metrics",
"interval": "5s"
}]
}
]
}
vim mysqld.json
{
"services": [
{
"id": "mysqld_exporter-node01",
"name": "node02",
"address": "192.168.136.160",
"port": 9104,
"tags": ["mysqld"],
"checks": [{
"http": "http://192.168.136.160:9104/metrics",
"interval": "5s"
}]
}
]
}
vim nginx.json
{
"services": [
{
"id": "nginx_exporter-node01",
"name": "node03",
"address": "192.168.136.160",
"port": 9913,
"tags": ["nginx"],
"checks": [{
"http": "http://192.168.136.160:9913/metrics",
"interval": "5s"
}]
}
]
}
(3)修改 prometheus 配置文件
vim /usr/local/prometheus/prometheus.yml
......
- job_name: nodes
consul_sd_configs: #指定使用 consul 服务发现
- server: 192.168.136.160:8500 #指定 consul 服务的端点列表
tags: #指定 consul 服务发现的 services 中哪些 service 能够加入到 prometheus 监控的标签
- nodes
refresh_interval: 1m
........
systemctl reload prometheus
浏览器查看 Prometheus 页面的 Status -> Targets
#让 consul 注销 Service
consul services deregister -id="node_exporter-node02"
#重新注册
consul services register /etc/consul/nodes.json
1.3 基于 Kubernetes API 的服务发现
基于 Kubernetes API 的服务发现机制,支持将API Server 中 Node、Service、Endpoint、Pod 和 Ingress 等资源类型下相应的各资源对象视作 target, 并持续监视相关资源的变动
●Node、Service、Endpoint、Pod 和 Ingress 资源分别由各自的发现机制进行定义
●负责发现每种类型资源对象的组件,在 Prometheus 中称为一个 role
●支持在集群上基于 DaemonSet 控制器部署 node-exporter 后发现各 Node 节点,也可以通过 kubelet 来作为 Prometheus 发现各 Node 节点的入口
#基于 Kubernetes 发现机制的部分配置参数
# The API server addresses. If left empty, Prometheus is assumed to run inside of the cluster and will discover API servers automatically
and use the pod's
# CA certificate and bearer token file at /var/run/secrets/kubernetes.io/serviceaccount/.
[ api_server: <host> ]
# The Kubernetes role of entities that should be discovered. One of endpoints, service, pod, node, or ingress.
role: <string>
# Optional authentication information used to authenticate to the API server.
# Note that 'basic_auth', 'bearer_token'和'bearer_token_file' 等认证方式互斥;
[ bearer_token: <secret> ]
[ bearer_token_file: <filename> ]
# TLS configuration.
tls_config:
# CA certificate to validate API server certificate with.
[ ca_file: <filename> ]
# Certificate and key files for client cert authentication to the server.
[ cert_file: <filename> ]
[ key_file: <filename> ]
# ServerName extension to indicate the name of the server.
[ server_name: <string> ]
# Optional namespace discovery. If omitted, all namespaces are used.
namespaces:
names:
[ - <string> ]
在K8S节点上
如果Prometheus是部署在k8s集群内部的,可以直接和apiserver通信。如果是集群外部的Prometheus可以通过以下步骤配置通过API Server实现服务发现。
RBAC授权
集群外部的Prometheus访问apiserver需要认证,还需要一定权限查询对应的资源对象。所以,这里先创建一个ServiceAccount和一个ClusterRole,并将它们绑定,之后集群外的Prometheus会以此ServiceAccount的身份访问apiserver。yaml文件内容如下:
apiVersion: v1
kind: Namespace
metadata:
name: monitoring
---
apiVersion: v1
kind: ServiceAccount
metadata:
name: outside-prometheus
namespace: monitoring
---
apiVersion: rbac.authorization.k8s.io/v1
kind: ClusterRole
metadata:
name: outside-prometheus
rules:
- apiGroups:
- ""
resources:
- nodes
- services
- endpoints
- pods
- nodes/proxy
verbs:
- get
- list
- watch
- apiGroups:
- "networking.k8s.io"
resources:
- ingresses
verbs:
- get
- list
- watch
- apiGroups:
- ""
resources:
- configmaps
- nodes/metrics
verbs:
- get
- nonResourceURLs:
- /metrics
verbs:
- get
---
apiVersion: rbac.authorization.k8s.io/v1
kind: ClusterRoleBinding
metadata:
name: outside-prometheus
roleRef:
apiGroup: rbac.authorization.k8s.io
kind: ClusterRole
name: outside-prometheus
subjects:
- kind: ServiceAccount
name: outside-prometheus
namespace: monitoring
获取ServiceAccount对应Secret资源对象中保存的token,然后将token保存到Prometheus节点上的文件里。
TOKEN=`kubectl get secret/$(kubectl -n monitoring get secret | awk '/outside-prometheus/{print $1}') -n monitoring -o jsonpath={.data.token} | base64 -d`
scp /etc/kubernetes/pki/ca.crt <prometheus_node>:/usr/local/prometheus/
在Prometheus 节点上
echo <token-value> > /usr/local/prometheus/kubernetes-api-token
cat /usr/local/prometheus/kubernetes-api-token #和上面获取的对比是否一致
修改Prometheus配置,添加job
#集群api-server自动发现job
- job_name: kubenetes-apiserver
kubernetes_sd_configs:
- role: endpoints
api_server: https://192.168.80.10:6443 #指定API Server地址
#这里的配置证书和token是连接API Server做服务发现时使用
tls_config:
ca_file: /usr/local/prometheus/pki/ca.crt #指定kubernetes ca根证书,用于验证api-server证书
# insecure_skip_verify: true #也可以使用此选项跳过证书验证
authorization:
credentials_file: /usr/local/prometheus/kubernetes-api-token #指定访问api-server时使用的token文件
scheme: https
#这里的配置证书和token是连接从api-server抓取数据时使用
tls_config:
ca_file: /usr/local/prometheus/pki/ca.crt
authorization:
credentials_file: /usr/local/prometheus/kubernetes-api-token
relabel_configs:
- source_labels: ["__meta_kubernetes_namespace", "__meta_kubernetes_endpoints_name", "__meta_kubernetes_endpoint_port_name"]
regex: default;kubernetes;https
action: keep
#集群节点自动发现job
- job_name: "kubernetes-nodes"
kubernetes_sd_configs:
- role: node #指定发现类型为node
api_server: https://192.168.80.10:6443
tls_config:
ca_file: /usr/local/prometheus/pki/ca.crt
authorization:
credentials_file: /usr/local/prometheus/kubernetes-api-token
relabel_configs:
- source_labels: ["__address__"] #重写target地址,默认端口是kubelet端口10250,修改为node-exporter端口9100
regex: (.*):10250
action: replace
target_label: __address__
replacement: $1:9100
- action: labelmap #保留之前存在的__meta_kubernetes_node_label开头的标签
regex: __meta_kubernetes_node_label_(.+)
验证
配置修改完成后,重载Prometheus,然后在界面查看target状态。
其实除了需要额外配置访问 API Server的证书外,其余配置和集群内的Prometheus服务发现配置基本一致。
另外,没有配置关于Pod的服务发现job,因为集群外的Prometheus无法访问集群内的Pod,需要添加路由规则才能实现互通。