80. 使用 grafana 和 prometheus监控Longhorn

SUSE Rancher Longhorn SUSE 牧场长角牛

Overview 概述

Longhorn integrates seamlessly with Prometheus, providing comprehensive visibility into the health and performance of your storage environment. Prometheus collects a wide range of metrics across Longhorn and related system components, enabling effective monitoring of volumes, disks, and nodes throughout the cluster.
Longhorn 与 Prometheus 无缝集成，全面了解您的存储环境健康状况和性能。Prometheus 收集了 Longhorn 及相关系统组件的广泛指标，能够有效监控集群中的卷、磁盘和节点。

Rancher integrates Prometheus and Grafana to monitor Kubernetes nodes, components, and workloads. To enable Rancher Monitoring refer the official documentation.
Rancher 集成了 Prometheus 和 Grafana，用于监控 Kubernetes 节点、组件和工作负载。 要启用牧场主监控，请参阅官方文档 。

If Rancher Monitoring is already enabled, Longhorn metrics can be scraped by deploying a ServiceMonitor in the Longhorn-system namespace.
如果已经启用了牧场主监控，可以通过在 Longhorn 系统命名空间中部署 ServiceMonitor 来抓取 Longhorn 的指标。

Step 1: Deploy Longhorn ServiceMonitor
步骤 1：部署 Longhorn ServiceMonitor

Apply the following manifest:
应用以下清单：

<span style="color:#000000"><span style="background-color:#ffffff"><span style="background-color:#efefef"><code>apiVersion: <a>After deployment, Prometheus will automatically discover and collect Longhorn metrics.
部署后，普罗米修斯会自动发现并收集长角牛的指标。
Validation: 验证：
Check metrics by running a query such as below in the Prometheus UI or check if the prometheus target is showing UP:
通过在 Prometheus 界面中运行查询，或检查 Prometheus 目标是否显示为 UP：
<span style="color:#000000"><span style="background-color:#ffffff"><span style="background-color:#efefef"><code>longhorn_disk_usage_bytes</code></span></span></span>
Step 2: Visualizing Metrics in Grafana
第二步：在 Grafana 中可视化度量
Accessing Grafana 访问格拉法纳
In Rancher UI, navigate to your Cluster
在 Rancher 界面中，导航到你的集群
Go to Cluster Tools (or Apps depending on Rancher version)
进入集群工具（或根据 Rancher 版本的应用）
SelectMonitoring
Click the Grafana icon/link to open in a new tab
点击 Grafana 图标/链接，在新标签页中打开
Log in (default credentials if unchanged):
登录（如果未更改，则使用默认凭证）：
Username:admin用户名：admin
Password:prom-operator
密码：prom-operator
Option A — Import a Prebuilt Longhorn Dashboard
选项 A — 导入预建的 Longhorn 仪表盘
Please refer to the community Grafana dashboard library for Longhorn.
请参阅社区 Grafana 仪表盘库中的 Longhorn。
Steps: 步骤：
Go to Dashboards → New → Import
前往仪表盘→新→导入
Select and import one of the available Longhorn dashboards
选择并导入其中一个可用的 Longhorn 仪表盘
Example: Dashboard ID 13032
示例：仪表盘 ID 13032
Select Prometheus as the data source
选择普罗米修斯作为数据源
This provides visualization of common Longhorn storage performance metrics.
这为常见的长角牛存储性能指标提供了可视化。
Option B — Create Custom Panels
选项 B — 创建自定义面板
In Grafana: From the sidebar, click → Dashboard → Add new dashboard → Add Visualization
在 Grafana 中：从侧边栏点击→仪表盘→添加新仪表盘→添加可视化
Select Prometheus as the data source
选择普罗米修斯作为数据源
You can either use the Query Builder to construct queries visually, or switch to the Code mode to write PromQL directly.
你可以使用查询构建器来可视化构建查询，或者切换到代码模式直接编写 PromQL。
Example PromQL for node storage utilization percentage:
节点存储利用率示例 PromQL：
<span style="color:#000000"><span style="background-color:#ffffff"><span style="background-color:#efefef"><code>(longhorn_node_storage_usage_bytes / longhorn_node_storage_capacity_bytes) * 100</code></span></span></span>
Additional recommended metrics:
其他推荐指标：
longhorn_node_count_total
longhorn_instance_manager_cpu_usage_millicpu
longhorn_node_cpu_capacity_millicpu
Adjust panel type, legends, and thresholds as required
根据需要调整面板类型、图例和阈值
Step 3: Save Dashboard 步骤3：保存仪表盘
Provide a descriptive name in title section and save the dashboard
在标题部分提供描述性名称并保存仪表盘
Optionally export as JSON for reuse across environments
可选地导出为 JSON 以便在不同环境中重复使用