prometheus监控磁盘容量（用prometheus监控GlusterFS）
prometheus监控磁盘容量（用prometheus监控GlusterFS）浏览器访问http://IP:9713/metrics或者服务器上curl下。[globals] gluster-mgmt = "glusterd" glusterd-dir = "/var/lib/glusterd" gluster-binary-path = "gluster" # If you want to connect to a remote gd1 host set the variable gd1-remote-host # However using a remote host restrict the gluster cli to read-only commands # The following collectors won't work in remote mode : gluster_volume_co
GlusterFS的测试环境搭建之后，在中间件的监控方面我首选prometheus，节约时间我们直接在github找已经造好的轮子。以下两个exporter值得推荐
1. https://github.com/gluster/gluster-prometheus # glusterfs官方仓库
2. https://github.com/ofesseler/gluster_exporter # 轻量简单
先测试下gluster-prometheus
安装1.github没有编译好的二进制文件（吐槽下，还要自己编译）
yum -y install go
export GOPATH=/root/go # 安装之后需要加下GOPATH环境变量
mkdir -p $GOPATH/src/github.com/gluster
cd $GOPATH/src/github.com/gluster
git clone https://github.com/gluster/gluster-prometheus.git
cd gluster-prometheus
# Install the required dependancies.
# Hint: assumes that GOPATH and PATH are already configured.
./scripts/install-reqs.sh
go get -u github.com/golang/dep/cmd/dep # 需要上网条件好
go get github.com/alecthomas/gometalinter # 需要上网条件好
PREFIX=/usr make # 需要上网条件好
PREFIX=/usr make install
按照github上教程，过程略坑，首先linux服务器上安装好go环境，以下步骤是需要注意的地方，有些步骤还需要上网条件好才行！！！我尝试使用GOPROXY没有成功。
yum -y install go # go环境准备
export GOPATH=/root/go
go get -u github.com/golang/dep/cmd/dep #go包管理器
go get github.com/alecthomas/gometalinter #go包管理器
编译完成得到可执行文件gluster-exporter，若读者朋友上网条件不太好，可以私信我，我会把编译好的二进制文件发给你。
2.执行PREFIX=/usr make install会产生systemd文件
install -D build/gluster-exporter /usr/sbin/gluster-exporter
install -D -m 0644 build/gluster-exporter.service /usr/lib/systemd/system/gluster-exporter.service
install -D -m 0600 ./extras/conf/gluster-exporter.toml.sample ""/etc/gluster-exporter/gluster-exporter.toml
systemctl enable gluster-exporter
systemctl start gluster-exporter
3.配置文件/etc/gluster-exporter/gluster-exporter.toml
[globals]
gluster-mgmt = "glusterd"
glusterd-dir = "/var/lib/glusterd"
gluster-binary-path = "gluster"
# If you want to connect to a remote gd1 host set the variable gd1-remote-host
# However using a remote host restrict the gluster cli to read-only commands
# The following collectors won't work in remote mode : gluster_volume_counts gluster_volume_profile
#gd1-remote-host = "localhost"
gd2-rest-endpoint = "http://127.0.0.1:24007"
port = 9713
metrics-path = "/metrics"
log-dir = "/var/log"
log-file = "gluster-exporter.log"
log-level = "info"
[collectors.gluster_ps]
name = "gluster_ps"
sync-interval = 5
disabled = false
[collectors.gluster_brick]
name = "gluster_brick"
sync-interval = 5
disabled = false
4.测试下是否获取到gluster数据
浏览器访问http://IP:9713/metrics或者服务器上curl下。
具体的metric的定义解释在https://github.com/gluster/gluster-prometheus/blob/master/docs/metrics.adoc，这个官方的exporter指标还是挺丰富的，gfs的CPU，内存，inode，读写，磁盘等信息都有，能满足监控的要求。
注意点glusterfs最好每个节点都跑下，有些集群的指标为了避免重复，所以first up peer in peer list will export these metrics，可以查看issues，https://github.com/gluster/gluster-prometheus/issues/148
我们再测下
github.com/ofesseler/gluster_exporter
安装go get github.com/ofesseler/gluster_exporter # golang.org可能需要网速快才行
./gluster_exporter
默认端口是9189
不需要配置文件，直接在对应的glusterfs节点上运行即可。
获取的指标相对官方来说少了一些，总体还是够用的，主要是通过以下命令获取到信息。
# gluster volume info
# gluster peer status
# gluster volume profile gv_test info cumulative
# gluster volume status all detail
添加到prometheus服务端
创建service和servicemonitor文件我们的prometheus是通过prometheus-operator跑在kubernetes集群上，所以还需要手动添加下service和servicemonitor
1.创建一个服务文件
apiVersion: v1
kind: Service
metadata:
name: gluster-exporter
namespace: monitoring
labels:
k8s-app: gluster-exporter
spec:
type: ClusterIP
clusterIP: None
ports:
- name: port
port: 9189
protocol: TCP
---
apiVersion: v1
kind: Endpoints
metadata:
name: glusterexporter
namespace: monitoring
labels:
k8s-app: glusterexporter
subsets:
- addresses:
- ip: 192.168.2.61
nodeName: gfs
- ip: 192.168.2.57
nodeName: node01
ports:
- name: port
port: 9189
protocol: TCP
2.创建一个servicemonitor文件用来做prometheus的target服务发现
apiVersion: monitoring.coreos.com/v1
kind: ServiceMonitor
metadata:
name: gluster-exporter
namespace: monitoring
labels:
k8s-app: gluster-exporter
spec:
endpoints:
- port: port
interval: 30s
scheme: http
selector:
matchLabels:
k8s-app: gluster-exporter
namespaceSelector:
matchNames:
- monitoringgrafana绘图
grafana的官方有对应的glusterfs的模板，喜欢的读者可以去下载
https://grafana.com/grafana/dashboards?search=gluster
我没有使用grafana上的模板，自己做了一个简单的，如下图：
