适用于 Azure 数据资源管理器的 Azure Monitor(预览版)Azure Monitor for Azure Data Explorer (preview)

适用于 Azure 数据资源管理器的 Azure Monitor(预览版)提供群集性能、操作、使用情况和故障的统一视图,可用于对群集进行全面的监视。Azure Monitor for Azure Data Explorer (preview) provides comprehensive monitoring of your clusters by delivering a unified view of your cluster performance, operations, usage, and failures. 本文将帮助你了解如何加入和使用适用于 Azure 数据资源管理器的 Azure Monitor(预览版)。This article will help you understand how to onboard and use Azure Monitor for Azure Data Explorer (preview).

适用于 Azure 数据资源管理器的 Azure Monitor(预览版)简介Introduction to Azure Monitor for Azure Data Explorer (preview)

在深入了解此体验之前,应该了解它如何呈现和可视化信息。Before jumping into the experience, you should understand how it presents and visualizes information.

  • 大规模透视图,显示群集主要指标的快照视图,以用于轻松跟踪查询、引入和导出操作的性能。At scale perspective showing a snapshot view of your clusters' primary metrics, to easily track performance of queries, ingestion, and export operations.
  • 特定 Azure 数据资源管理器群集的 深入分析,可帮助执行详细分析。Drill down analysis of a particular Azure Data Explorer cluster to help perform detailed analysis.
  • 可自定义:你可以更改要查看、修改的指标,或设置与限制相符的阈值,并保存你自己的自定义工作簿。Customizable where you can change which metrics you want to see, modify or set thresholds that align with your limits, and save your own custom workbooks. 可以将工作簿中的图表固定到 Azure 面板。Charts in the workbook can be pinned to Azure dashboards.

Azure Monitor 中的视图(大规模透视图)View from Azure Monitor (at scale perspective)

在 Azure Monitor 中,可以查看群集的主要性能指标(包括订阅中多个群集的查询、引入和导出操作的指标),并帮助识别性能问题。From Azure Monitor, you can view the main performance metrics for the cluster, including metrics for queries, ingestion, and export operations from multiple clusters in your subscription, and help identify performance problems.

若要查看所有订阅中群集的性能,请执行以下步骤:To view the performance of your clusters across all your subscriptions, perform the following steps:

  1. 登录到 Azure 门户Sign into the Azure portal

  2. 在 Azure 门户的左侧窗格中选择“监视”,然后在“见解中心”部分下选择“Azure 数据资源管理器群集(预览版)”。 Select Monitor from the left-hand pane in the Azure portal, and under the Insights Hub section, select Azure Data Explorer Clusters (preview).

“概览”体验的屏幕截图,其中显示了多个关系图

概述选项卡Overview tab

在所选订阅的“概述”选项卡上,表格显示了订阅中已分组的 Azure 数据资源管理器群集的交互指标。On the Overview tab for the selected subscription, the table displays interactive metrics for the Azure Data Explorer clusters grouped within the subscription. 你可以根据从下面的下拉列表中选择的选项来筛选结果:You can filter results based on the options you select from the following drop-down lists:

  • 订阅 - 仅列出了包含 Azure 数据资源管理器群集的订阅。Subscriptions - only subscriptions that have Azure Data Explorer clusters are listed.

  • Azure 数据资源管理器群集 - 默认情况下,最多只会预先选择 5 个群集。Azure Data Explorer clusters - by default, only up to five clusters are pre-selected. 如果在范围选择器中选择所有或多个群集,则最多会返回 200 个群集。If you select all or multiple clusters in the scope selector, up to 200 clusters will be returned.

  • 时间范围 - 系统默认会根据你所做的选择显示过去 24 小时的相应信息。Time Range - by default, displays the last 24 hours of information based on the corresponding selections made.

下拉列表下的计数器磁贴汇总了所选订阅中 Azure 数据资源管理器群集的总数,并反映了选择的群集数。The counter tile, under the drop-down list, rolls-up the total number of Azure Data Explorer clusters in the selected subscriptions and reflects how many are selected. 以下各列有条件颜色编码:“保持活动状态”、“CPU”、“引入利用率”和“缓存利用率”。There are conditional color-codings for the columns: Keep alive, CPU, Ingestion Utilization, and Cache Utilization. 以橙色编码的单元格中的值对于群集而言是不可持续的。Orange-coded cells have values that are not sustainable for the cluster.

为了更好地了解每个指标的含义,建议通读有关 Azure 数据资源管理器指标的文档。To better understand what each of the metrics represent, we recommend reading through the documentation on Azure Data Explorer metrics.

“查询性能”选项卡Query Performance tab

此选项卡显示查询持续时间、并发查询总数和受限制查询总数。This tab shows the query duration, total number of concurrent queries, and the total number of throttled queries.

“查询性能”选项卡的屏幕截图

“引入性能”选项卡Ingestion Performance tab

此选项卡显示引入延迟、成功的引入结果数、失败的引入结果数、引入量,以及事件中心/IoT 中心处理的事件数。This tab shows the ingestion latency, succeeded ingestion results, failed ingestion results, ingestion volume, and events processed for Event/IoT Hubs.

“引入性能”选项卡的屏幕截图Screenshot of ingestion performance tab

“流引入性能”选项卡Streaming Ingest Performance tab

此选项卡提供有关平均数据速率、平均持续时间和请求速率的信息。This tab provides information on the average data rate, average duration, and request rate.

“导出性能”选项卡Export Performance tab

此选项卡提供有关连续导出操作的已导出记录数、延迟、挂起计数和利用率百分比的信息。This tab provides information on exported records, lateness, pending count, and utilization percentage for continuous export operations.

来自 Azure 数据资源管理器群集资源的视图(深入分析)View from an Azure Data Explorer Cluster resource (drill down analysis)

若要直接从 Azure 数据资源管理器群集访问适用于 Azure 数据资源管理器群集的 Azure Monitor,请执行以下操作:To access Azure Monitor for Azure Data Explorer Clusters directly from an Azure Data Explorer Cluster:

  1. 在 Azure 门户中,选择“Azure 数据资源管理器群集”。In the Azure portal, select Azure Data Explorer Clusters.

  2. 从列表中选择一个 Azure 数据资源管理器群集。From the list, choose an Azure Data Explorer Cluster. 在监视部分选择“见解(预览版)”。In the monitoring section, choose Insights (preview).

还可以通过从 Azure Monitor 见解视图中选择某个 Azure 数据资源管理器群集的资源名称来访问这些视图。These views are also accessible by selecting the resource name of an Azure Data Explorer cluster from within the Azure Monitor insights view.

适用于 Azure 数据资源管理器的 Azure Monitor 将日志与指标结合起来,以提供全局监视解决方案。Azure Monitor for Azure Data Explorer combines both logs and metrics to provide a global monitoring solution. 包含基于日志的可视化效果需要用户启用其 Azure 数据资源管理器群集的诊断日志记录,并将日志发送到 Log Analytics 工作区The inclusion of logs-based visualizations requires users to enable diagnostic logging of their Azure Data Explorer cluster and send them to a Log Analytics workspace.. 应启用的诊断日志包括:CommandQueryTableDetailsTableUsageStatisticsThe diagnostic logs that should be enabled are: Command, Query, TableDetails, and TableUsageStatistics.

显示文本“启用日志以进行监视”的蓝色按钮屏幕截图

“概述”选项卡显示:The Overview tab shows:

  • “指标”磁贴,其中突出显示了群集的可用性和总体状态,以用于快速评估其运行状况。Metrics tiles highlighting the availability and overall status of the cluster to quickly assess its health.

  • 有效顾问建议和资源运行状况的摘要。A summary of active Advisor recommendations and resource health status.

  • 一些图表,其中显示了一段时间内 CPU 和内存消耗量最大的几个消耗者以及唯一用户数。Charts showing the top CPU and memory consumers and the number of unique users over time.

来自 Azure 数据资源管理器群集资源的视图的屏幕截图Screenshot of view from an Azure Data Explorer cluster resource

“关键指标”选项卡显示某些已分组的群集指标的统一视图,这些指标分组为:常规指标、查询相关、引入相关,以及流引入相关的指标。The Key Metrics tab shows a unified view of some of the cluster's metrics, grouped by: general metrics, query-related, ingestion-related, and streaming ingestion-related metrics.

“故障”视图的屏幕截图Screenshot of failures view

“使用情况”选项卡可让用户深入了解群集命令和查询的性能。The Usage tab allows users to deep dive into the performance of the cluster's commands and queries. 在此页中,可以:On this page, you can:

  • 查看哪些工作负载组、用户和应用程序发送的查询最多或者消耗的 CPU 和内存最多(这样就可以了解哪些工作负载正在提交资源消耗量最大的查询供群集处理)。See which workload groups, users and applications are sending the most queries or consuming the most CPU and memory (so you can understand which workloads are submitting the heaviest queries for the cluster to process).
  • 按照失败的查询数识别排名靠前的工作负载组、用户和应用程序。Identify top workload groups, users and applications by failed queries.
  • 通过将查询数与各个工作负载组、用户和应用程序(在过去 16 天)的历史每日平均值进行比较,识别查询数最近发生的变化。Identify recent changes in the number of queries, compared to the historical daily average (over the past 16 days), by workload group, user and application.
  • 按照工作负载组、用户、应用程序和命令类型,识别查询数、内存和 CPU 消耗量的趋势与峰值。Identify trends and peaks in the number of queries, memory, and CPU consumption by workload group, user, application and command type.

操作视图的屏幕截图,其中的圆环图按命令和查询计数显示了排名最高的应用程序、按命令和查询计数显示了排名最高的主体,并且按命令类型显示了排名最高的命令Screenshot of operations view with donut charts of top application by command and query count, top principals by command and query count, and top commands by command types

操作视图的屏幕截图,其中的折线图按应用程序显示了查询计数、按应用程序显示了内存总消耗量,并且按应用程序显示了 CPU 总消耗量Screenshot of operations view with line charts of query count by application, total memory by application and total CPU by application

“表”选项卡显示群集中表的最新属性和历史属性。The tables tab shows the latest and historical properties of tables in the cluster. 可以查看哪些表占用的空间最多,按表大小、热数据和一段时间的行数跟踪增长历史记录。You can see which tables are consuming the most space, track growth history by table size, hot data, and the number of rows over time.

“缓存”选项卡可让用户分析其实际查询的回溯模式,并将其与配置的缓存策略(针对每个表)进行比较。The cache tab allows users to analyze their actual queries' look back patterns and compare them to the configured cache policy (for each table). 可以识别大多数查询使用的表和根本未被查询的表,并相应地调整缓存策略。You can identify tables used by the most queries and tables that are not queried at all, and adapt the cache policy accordingly. Azure 顾问中的缓存缩减建议适用于“受数据限制”的群集(即,这些群集的 CPU 利用率和引入利用率较低,但由于数据容量较高,群集无法横向或纵向缩减)。Cache reduction recommendations in Azure Advisor are available for clusters that are "bounded by data" (meaning the cluster has low CPU and low ingestion utilization, but because of high data capacity, the cluster could not scale-in or scale-down).

缓存详细信息的屏幕截图Screenshot of cache details

固定到 Azure 仪表板Pin to Azure dashboard

可将(“大规模”透视图的)任一指标部分固定到 Azure 仪表板,只需选择该部分右上角的图钉图标即可。You can pin any one of the metric sections (of the "at-scale" perspective) to an Azure dashboard by selecting the pushpin icon at the top right of the section.

已选择的固定图标的屏幕截图

自定义适用于 Azure 数据资源管理器群集的 Azure MonitorCustomize Azure Monitor for Azure Data Explorer Cluster

此部分重点介绍了编辑工作簿的常见方案,以进行自定义来满足数据分析需求:This section highlights common scenarios for editing the workbook to customize in support of your data analytics needs:

  • 将工作簿限定为始终选择特定的订阅或 Azure 数据资源管理器群集Scope the workbook to always select a particular subscription or Azure Data Explorer Cluster(s)
  • 更改网格中的指标Change metrics in the grid
  • 更改阈值或颜色呈现/编码Change thresholds or color rendering/coding

从顶部工具栏选择“自定义”按钮可启用编辑模式,以开始自定义。You can begin customizations by enabling the editing mode, by selecting the Customize button from the top toolbar.

“自定义”按钮的屏幕截图

自定义项保存到自定义工作簿,以防止覆盖已发布工作簿中的默认配置。Customizations are saved to a custom workbook to prevent overwriting the default configuration in our published workbook. 无论是在专用的“我的报表”部分,还是在有权访问资源组的任何用户均可访问的“共享报表”部分,工作簿都保存在某个资源组中。Workbooks are saved within a resource group, either in the My Reports section that is private to you or in the Shared Reports section that's accessible to everyone with access to the resource group. 保存自定义工作簿后,需要转到工作簿库来启动它。After you save the custom workbook, you need to go to the workbook gallery to launch it.

工作簿库的屏幕截图

故障排除Troubleshooting

如需常规故障排除指南,请参阅专用的基于工作簿的见解故障排除文章For general troubleshooting guidance, refer to the dedicated workbook-based insights troubleshooting article.

本部分将帮助你诊断和排查在使用适用于 Azure 数据资源管理器群集的 Azure Monitor(预览版)时可能会遇到的一些常见问题。This section will help you with the diagnosis and troubleshooting of some of the common issues you may encounter when using Azure Monitor for Azure Data Explorer Cluster (preview). 使用下面的列表来查找与具体问题相关的信息。Use the list below to locate the information relevant to your specific issue.

为什么在订阅选取器中看不到所有订阅?Why don't I see all my subscriptions in the subscription picker?

我们只显示从所选订阅筛选器中选择的包含 Azure 数据资源管理器群集的订阅(在 Azure 门户标题上的“目录 + 订阅”中选择)。We only show subscriptions that contain Azure Data Explorer Clusters, chosen from the selected subscription filter, which are selected in the "Directory + Subscription" in the Azure portal header.

订阅筛选器的屏幕截图

“使用情况”、“表”或“缓存”部分下为何不显示我的 Azure 数据资源管理器群集的任何数据?Why do I not see any data for my Azure Data Explorer Cluster under the Usage, Tables or Cache sections?

若要查看基于日志的数据,需要为你要监视的每个 Azure 数据资源管理器群集启用诊断日志To view your logs-based data, you will need to enable diagnostic logs for each of the Azure Data Explorer Clusters you want to monitor. 可以在每个群集的诊断设置下完成此操作。This can be done under the diagnostic settings for each cluster. 需要将数据发送到 Log Analytics 工作区。You will need to send your data to a Log Analytics workspace. 应启用的诊断日志包括:Command、Query、TableDetails 和 TableUsageStatistics。The diagnostic logs that should be enabled are: Command, Query, TableDetails, and TableUsageStatistics.

我已经为我的 Azure 数据资源管理器群集启用了日志,但为何“命令”和“查询”下仍不显示我的数据?I have already enabled logs for my Azure Data Explorer Cluster, why am I still unable to see my data under Commands and Queries?

目前,诊断日志无法以追溯方式工作,只有在对 Azure 数据资源管理器执行了操作后,才会开始显示数据。Currently, diagnostic logs do not work retroactively, so the data will only start appearing once there have been actions taken to your Azure Data Explorer. 因此,可能需要在一段时间后才会显示你的数据,所需时间为几小时到一天不等,具体取决于 Azure 数据资源管理器群集的活跃程度。Therefore, it may take some time, ranging from hours to a day, depending on how active your Azure Data Explorer cluster is.

后续步骤Next steps

查看使用 Azure Monitor 工作簿创建交互式报表,了解工作簿旨在支持的方案、创作新报表和自定义现有报表的方式,以及更多信息。Learn the scenarios workbooks are designed to support, how to author new and customize existing reports, and more by reviewing Create interactive reports with Azure Monitor workbooks.