什么是 Azure 时序见解 Gen1?What is Azure Time Series Insights Gen1?


这是一篇 Gen1 文章。This is a Gen1 article.

Azure 时序见解用于存储、可视化和查询大量时序数据(例如 IoT 设备所生成的数据)。Azure Time Series Insights is built to store, visualize, and query large amounts of time series data, such as that generated by IoT devices. 如果你想要在云中存储、管理、查询或可视化时序数据,则 Azure 时序见解可能会很适合你。If you want to store, manage, query, or visualize time series data in the cloud, Azure Time Series Insights is likely right for you.

Azure 时序见解流程图Azure Time Series Insights flowchart

Azure 时序见解包含四个关键作业:Azure Time Series Insights has four key jobs:

  • 它与 Azure IoT 中心和 Azure 事件中心等云网关完全集成。It's fully integrated with cloud gateways like Azure IoT Hub and Azure Event Hubs. 它可以轻松连接到这些事件源,并根据消息和结构(在干净的行和列中包含数据)分析 JSON。It easily connects to these event sources and parses JSON from messages and structures that have data in clean rows and columns. 它将元数据与遥测数据联接,并在纵栏式存储中对数据编制索引。It joins metadata with telemetry and indexes your data in a columnar store.
  • Azure 时序见解管理数据存储。Azure Time Series Insights manages the storage of your data. 为确保始终能够轻松访问数据,它将数据存储在内存和 SSD 中长达 400 天。To make sure that data is always easily accessible, it stores your data in memory and SSDs for up to 400 days. 可以在数秒内以交互方式按需查询数十亿的事件。You can interactively query billions of events in seconds-on demand.
  • Azure 时序见解通过 Azure 时序见解资源管理器提供现成可用的可视化效果。Azure Time Series Insights provides out-of-the-box visualization through the Azure Time Series Insights Explorer.
  • Azure 时序见解通过两种途径提供查询服务,一种是使用 Azure 时序见解资源管理器,另一种是使用可轻松集成的 API 将时序数据嵌入自定义应用程序中。Azure Time Series Insights provides a query service, both in the Azure Time Series Insights Explorer and by using APIs that are easy to integrate to embed your time series data into custom applications.

如果构建供内部使用或供外部客户使用的应用程序,则可将 Azure 时序见解用作后端。If you build an application for internal consumption or for external customers to use, you can use Azure Time Series Insights as a back end. 可以使用它来索引、存储和聚合时序数据。You can use it to index, store, and aggregate time series data. 若要在此之上构建自定义可视化效果和用户体验,请使用客户端 SDKTo build a custom visualization and user experience on top, use the Client SDK. Azure 时序见解还配备了多个查询 API 以启用这些自定义方案。Azure Time Series Insights is also equipped with several Query APIs to enable these customized scenarios.

时序数据表示资产或过程是如何随时间变化的。Time series data represents how an asset or process changes over time. 时序数据按时间戳进行索引,时间是用于组织此类数据的最有意义的轴。Time series data is indexed by timestamps, and time is the most meaningful axis along which such data is organized. 时序数据通常按顺序到达,因此被视为插入,而不是数据库的更新。Time series data typically arrives in sequential order, so it's treated as an insert rather than an update to your database.

在大型卷中存储、索引、查询、分析和可视化时序数据可能很具挑战性。It can be a challenge to store, index, query, analyze, and visualize time series data in large volumes. Azure 时序见解会捕获每一个新事件并将其存储为一行,所做的更改可以在一段时间内进行有效的度量。Azure Time Series Insights captures and stores every new event as a row, and change is efficiently measured over time. 因此,可以通过回顾过去的见解来预测未来的更改。As a result, you can look backwards to draw insights from the past to help predict future change.

主要方案Primary scenarios

  • 以可缩放的方式存储时序数据。Store time series data in a scalable way.

    在其核心,Azure 时序见解有一个数据库,是专门围绕时序数据而设计的。At its core, Azure Time Series Insights has a database designed with time series data in mind. 因为是可缩放的和完全托管,所以 Azure 时序见解负责存储和管理事件。Because it's scalable and fully managed, Azure Time Series Insights handles the work of storing and managing events.

  • 以近实时方式浏览数据。Explore data in near real time.

    Azure 时序见解提供了一个资源管理器,用于可视化所有流式传输到环境中的数据。Azure Time Series Insights provides an Explorer that visualizes all data that streams into an environment. 连接到事件源后不久,便可在 Azure 时序见解中查看、浏览和查询事件数据。Shortly after you connect to an event source, you can view, explore, and query event data within Azure Time Series Insights. 数据可用于验证设备是否按预期方式发出数据,并监视 IoT 资产的运行状况、生产效率和总体效率。The data helps you to validate whether a device emits data as expected and to monitor an IoT asset for health, productivity, and overall effectiveness.

  • 进行根本原因分析并检测异常情况Perform root-cause analysis and detect anomalies.

    Azure 时序见解具有模式和透视视图等工具,用于执行和保存多步骤根本原因分析。Azure Time Series Insights has tools like patterns and perspective views to conduct and save multistep root-cause analysis. Azure 时序见解还可与 Azure 流分析等警报服务结合使用,让你能够在 Azure 时序见解资源管理器中近实时地查看警报和检测到的异常情况。Azure Time Series Insights also works with alerting services like Azure Stream Analytics so that you can view alerts and detected anomalies in near real time in the Azure Time Series Insights Explorer.

  • 获取从不同位置流式传输的时序数据的全局视图,进行多资产或多站点比较。Gain a global view of time series data that streams from disparate locations for multi-asset or site comparison.

    可以将多个事件源连接到 Azure 时序见解环境。You can connect multiple event sources to an Azure Time Series Insights environment. 这样就可以近实时地查看从多个不同位置一起流式传输进来的数据。This way you can view data that streams in from multiple, disparate locations together in near real time. 用户可以利用这种可见性与业务领导者共享数据。Users can take advantage of this visibility to share data with business leaders. 用户还可以与领域专家更好地协作,而这些专家则可以使用其专业技能帮助用户解决问题、应用最佳做法和共享知识。They can collaborate better with domain experts who can apply their expertise to help solve problems, apply best practices, and share learnings.

  • 在 Azure 时序见解的基础上构建客户应用程序。Build a customer application on top of Azure Time Series Insights.

    Azure 时序见解会公开 REST 查询 API,这些 API 可以用来构建使用时序数据的应用程序。Azure Time Series Insights exposes REST Query APIs that you can use to build applications that use time series data.


  • 快速入门:Azure 时序见解不需前期数据准备,因此你可以在 IoT 中心或事件中心快速连接到数百万事件。Get started quickly: Azure Time Series Insights doesn't require upfront data preparation, so you can quickly connect to millions of events in your IoT hub or event hub. 连接以后,即可可视化传感器数据并与之交互,快速验证 IoT 解决方案。After you connect, you can visualize and interact with sensor data to quickly validate your IoT solutions. 无需编写代码即可与数据交互,也无需学习新的语言。You can interact with your data without writing code, and you don't need to learn a new language. Azure 时序见解既为高级用户提供精细的自由文本查询图面,又提供点击浏览体验。Azure Time Series Insights provides a granular, free-text query surface for advanced users, and point-and-click exploration.

  • 近乎实时的见解:Azure 时序见解每天可以引入数百万个传感器事件,只有一分钟的延迟。Near real-time insights: Azure Time Series Insights can ingest millions of sensor events per day, with one-minute latency. 可以通过 Azure 时序见解深入分析传感器数据。Azure Time Series Insights helps you gain insights into your sensor data. 可以通过它来发现趋势和异常情况,进行根本原因分析并避免成本高昂的停机。Use it to spot trends and anomalies, conduct root-cause analyses, and avoid costly downtime. 在实时数据和历史数据之间进行交叉关联,帮助你发现数据中隐藏的趋势。Cross-correlation between real-time and historical data helps you find hidden trends in the data.

  • 生成自定义解决方案:将 Azure 时序见解数据嵌入现有应用程序。Build custom solutions: Embed Azure Time Series Insights data into your existing applications. 也可以通过 Azure 时序见解 REST API 创建新的自定义解决方案。You also can create new custom solutions with the Azure Time Series Insights REST APIs. 创建个性化视图,以便通过这种共享方便他人浏览你的见解。Create personalized views you can share for others to explore your insights.

  • 可伸缩性:Azure 时序见解旨在支持大规模 IoT。Scalability: Azure Time Series Insights is designed to support IoT at scale. 时序见解每天可以引入 100 万到 1 亿个事件,默认保留时间为 31 天。It can ingress from 1 million to 100 million events per day, with a default retention span of 31 days. 可以通过近实时方式可视化和分析实时数据流以及历史数据。You can visualize and analyze live data streams in near real time, alongside historical data.

入门Get started

请按照以下步骤开始。To get started, follow these steps.

  1. 在 Azure 门户中预配 Azure 时序见解环境。Provision an Azure Time Series Insights environment in the Azure portal.
  2. 连接到 IoT 中心或事件中心等事件源。Connect to an event source like an IoT hub or an event hub.
  3. 上传参考数据。Upload reference data. 这不是一项附加服务。This isn't an additional service.
  4. 使用 Azure 时序见解资源管理器在几分钟内查看并显示数据。Review and display your data in minutes with the Azure Time Series Insights Explorer.


使用 Azure 时序见解资源管理器查看、分析和发现数据趋势。View, analyze, and discover trends in your data using the Azure Time Series Insights Explorer.

Azure 时序见解资源管理器

了解如何使用 Azure 时序见解资源管理器获取数据趋势。Learn how to use the Azure Time Series Insights Explorer and draw insights from your data.

后续步骤Next steps