使用 Azure 顾问提高 Azure 应用程序的性能Improve the performance of Azure applications by using Azure Advisor

Azure 顾问中的性能建议有助于提高业务关键型应用程序的速度和响应能力。The performance recommendations in Azure Advisor can help improve the speed and responsiveness of your business-critical applications. 可通过顾问从顾问仪表板的“性能”选项卡获取性能建议。You can get performance recommendations from Advisor on the Performance tab of the Advisor dashboard.

缩短流量管理器配置文件上的 DNS 生存时间,以更快地故障转移到正常运行的终结点Reduce DNS time-to-live on your Traffic Manager profile to fail over to healthy endpoints faster

可以使用 Azure 流量管理器配置文件中的生存时间 (TTL) 设置指定在给定终结点停止响应查询时切换终结点的速度。You can use time-to-live (TTL) settings on your Azure Traffic Manager profile to specify how quickly to switch endpoints if a given endpoint stops responding to queries. 如果降低 TTL 值,则客户端将更快地路由到正常运行的终结点。If you reduce the TTL values, clients will be routed to functioning endpoints faster.

Azure 顾问可识别配置了较长 TTL 的流量管理器配置文件。Azure Advisor identifies Traffic Manager profiles that have a longer TTL configured. 它会建议将 TTL 配置为 20 秒或 60 秒,具体取决于配置文件是否配置为使用快速故障转移It recommends configuring the TTL to either 20 seconds or 60 seconds, depending on whether the profile is configured for Fast Failover.

使用 SQL 数据库顾问(暂时禁用)提高数据库性能Improve database performance by using SQL Database Advisor (temporarily disabled)

Azure 顾问针对所有 Azure 资源提供一个一致且统一的建议视图。Azure Advisor provides a consistent, consolidated view of recommendations for all your Azure resources. 它与 SQL 数据库顾问集成,为你提供用于提高数据库性能的建议。It integrates with SQL Database Advisor to bring you recommendations for improving the performance of your databases. SQL 数据库顾问通过分析使用情况历史记录来评估数据库的性能, SQL Database Advisor assesses the performance of your databases by analyzing your usage history. 然后提供最适合运行数据库典型工作负荷的建议。It then offers recommendations that are best suited for running the database's typical workload.

备注

需要先使用数据库大约一周时间,并在这一周内完成一些一致的活动,然后才能获取建议。Before you can get recommendations, your database needs to be in use for about a week, and within that week there needs to be some consistent activity. SQL 数据库顾问优化一致的查询模式比优化随机的突发活动更加轻松。SQL Database Advisor can optimize more easily for consistent query patterns than for random bursts of activity.

有关详细信息,请参阅 SQL 数据库顾问For more information, see SQL Database Advisor.

将存储客户端库升级到最新版本,以提高可靠性和性能Upgrade your Storage client library to the latest version for better reliability and performance

最新版存储客户端库 SDK 包含修补程序,可解决由客户报告的问题以及通过我们的 QA 过程主动识别的问题。The latest version of the Storage client library SDK contains fixes to problems reported by customers and proactively identified through our QA process. 除了新功能,最新版本还提供可靠性和性能优化,可以改进你在使用 Azure 存储时的整体体验。The latest version also carries reliability and performance optimization together with new features that can improve your overall experience with using Azure Storage. 如果使用的是旧版本,则顾问会提供升级到最新版本的 SDK 所需的建议和步骤。Advisor provides recommendations and steps needed to upgrade to the latest version of the SDK if you're using a stale version. 这些建议适用于受支持的语言:C++ 和 .NET。The recommendations are for supported languages: C++ and .NET.

提高应用服务性能和可靠性Improve App Service performance and reliability

Azure 顾问集成用于改善应用服务体验以及发现相关平台功能的建议。Azure Advisor integrates recommendations for improving your App Service experience and discovering relevant platform capabilities. 以下为应用服务建议示例:Examples of App Service recommendations are:

  • 检测其中的应用运行时已耗尽内存或 CPU 资源的实例,并提出缓解措施。Detection of instances where memory or CPU resources are exhausted by app runtimes, with mitigation options.
  • 检测在其中归置资源(如 Web 应用和数据库)可提高性能并降低成本的实例。Detection of instances where co-locating resources like web apps and databases can improve performance and reduce cost.

有关详细信息,请参阅 Azure 应用服务最佳做法For more information, see Best practices for Azure App Service.

使用托管磁盘防止磁盘 I/O 限制Use managed disks to prevent disk I/O throttling

顾问可识别属于即将达到其可伸缩性目标的存储帐户的虚拟机。Advisor identifies virtual machines that belong to a storage account that's reaching its scalability target. 这种情况使得这些 VM 容易遭受 I/O 限制。This condition makes those VMs susceptible to I/O throttling. 顾问会建议它们使用托管磁盘来防止性能降低。Advisor will recommend that they use managed disks to prevent performance degradation.

可以通过使用高级磁盘来提高虚拟机磁盘的性能和可靠性Improve the performance and reliability of virtual machine disks by using Premium Storage

顾问可识别标准磁盘在存储帐户上有大量事务的虚拟机。Advisor identifies virtual machines with standard disks that have a high volume of transactions on your storage account. 它会建议升级到高级磁盘。It recommends upgrading to premium disks.

Azure 高级存储为运行 I/O 密集型工作负荷的虚拟机提供高性能、低延迟的磁盘支持。Azure Premium Storage delivers high-performance, low-latency disk support for virtual machines that run I/O-intensive workloads. 使用高级存储帐户的虚拟机磁盘在固态硬盘 (SSD) 上存储数据。Virtual machine disks that use Premium Storage accounts store data on solid-state drives (SSDs). 若要让应用程序实现最佳性能,建议将任何需要高 IOPS 的虚拟机磁盘迁移到高级存储。For the best performance for your application, we recommend that you migrate any virtual machine disks that require high IOPS to Premium Storage.

删除 Azure Synapse Analytics 表上的数据倾斜以提高查询性能Remove data skew on your Azure Synapse Analytics tables to increase query performance

运行工作负荷时,数据倾斜可能会造成不必要的数据移动或资源瓶颈。Data skew can cause unnecessary data movement or resource bottlenecks when you run your workload. 顾问可检测大于 15% 的分布数据倾斜。Advisor detects distribution data skew of greater than 15%. 建议重新分布数据并重新考虑表分布键选项。It recommends that you redistribute your data and revisit your table distribution key selections. 若要了解有关识别和删除倾斜的详细信息,请参阅倾斜疑难解答To learn more about identifying and removing skew, see troubleshooting skew.

在 Azure Synapse Analytics 表中创建或更新过期表统计信息,以提高查询性能Create or update outdated table statistics in your Azure Synapse Analytics tables to increase query performance

顾问可识别不包含最新表统计信息的表,并建议创建或更新统计信息。Advisor identifies tables that don't have up-to-date table statistics and recommends creating or updating the statistics. Azure Synapse Analytics 中的查询优化器使用最新统计信息来估算查询结果的基数或行数。The query optimizer in Azure Synapse Analytics uses up-to-date statistics to estimate the cardinality or number of rows in query results. 借助这些估算值,查询优化器可创建查询计划以提供最快的性能。These estimates enable the query optimizer to create a query plan to provide the fastest performance.

改进 MySQL 连接管理Improve MySQL connection management

顾问分析可以指示连接到 MySQL 服务器的应用程序可能无法高效管理连接。Advisor analysis can indicate that your application connecting to a MySQL server might not be managing connections efficiently. 这种情况可能导致不必要的资源消耗和应用程序的总体延迟过高。This condition could lead to unnecessary resource consumption and overall higher application latency. 为了改进连接管理,建议你减少短期连接数并消除不必要的空闲连接。To improve connection management, we recommend that you reduce the number of short-lived connections and eliminate unnecessary idle connections. 可以通过配置服务器端连接池(如 ProxySQL)来进行这些改进。You can make these improvements by configuring a server-side connection pooler, like ProxySQL.

将当前的 Compute Management SDK 版本更新到最新版本Update your current Compute Management SDK version to the most recent version

顾问可识别具有使用过期 Compute Management SDK 版本的操作的订阅。Advisor identifies subscriptions which have operations using outdated Compute Management SDK versions. 这可能会影响工作负荷的安全性和性能,因此,顾问建议切换到最新版本的 Compute Management SDK。This might impact the security and performance of your workloads and thus, Advisor recommends you to switch to the latest version of Compute Management SDK.

通过纵向扩展来优化 Azure Synapse Analytics 表上的缓存利用率,从而提高查询性能Scale up to optimize cache utilization on your Azure Synapse Analytics tables to increase query performance

Azure 顾问可检测 Azure Synapse Analytics 表是否具有较高的缓存利用率和较低的命中率。Azure Advisor detects whether your Azure Synapse Analytics tables have a high cache-used percentage and a low hit percentage. 这种情况表示存在高缓存逐出,这可能影响 Azure Synapse Analytics 实例的性能。This condition indicates high cache eviction, which can affect the performance of your Azure Synapse Analytics instance. 顾问建议纵向扩展 Azure Synapse Analytics 实例,以确保为工作负荷分配足够的缓存容量。Advisor recommends that you scale up your Azure Synapse Analytics instance to ensure you allocate enough cache capacity for your workload.

将 Azure Synapse Analytics 表转换为复制表,以提高查询性能Convert Azure Synapse Analytics tables to replicated tables to increase query performance

顾问可识别不是复制表,但将从转换中受益的表。Advisor identifies tables that aren't replicated tables but that would benefit from conversion. 它会建议转换这些表。It suggests that you convert these tables. 建议基于:Recommendations are based on:

  • 复制表的大小。The size of the replicated table.
  • 列数。The number of columns.
  • 表分布类型。The table distribution type.
  • Azure Synapse Analytics 表中的分区数。The number of partitions on the Azure Synapse Analytics table.

还可能会根据上下文在建议中提供额外的启发。Additional heuristics might be provided in the recommendation for context. 若要详细了解此建议是如何确定的,请参阅 Azure Synapse Analytics 建议To learn more about how this recommendation is determined, see Azure Synapse Analytics recommendations.

将存储帐户迁移到 Azure 资源管理器来获取最新 Azure 功能Migrate your storage account to Azure Resource Manager to get the latest Azure features

将存储帐户部署模型迁移到 Azure 资源管理器,以利用:Migrate your storage account deployment model to Azure Resource Manager to take advantage of:

  • 模板部署。Template deployments.
  • 其他安全选项。Additional security options.
  • 升级到 GPv2 帐户的能力,因此可以使用最新的 Azure 存储功能。The ability to upgrade to a GPv2 account so you can use the latest Azure Storage features.

顾问可识别使用经典部署模型的任何独立存储帐户,并建议迁移到资源管理器部署模型。Advisor identifies any stand-alone storage accounts that are using the classic deployment model and recommends migrating to the Resource Manager deployment model.

备注

Azure Monitor 中的经典警报已于 2019 年 8 月停用。Classic alerts in Azure Monitor were retired in August 2019. 建议升级经典存储帐户以使用资源管理器在新平台中保留警报功能。We recommended that you upgrade your classic storage account to use Resource Manager to retain alerting functionality with the new platform.

设计存储帐户,防止达到最大订阅数限制Design your storage accounts to prevent reaching the maximum subscription limit

一个 Azure 区域最多支持每个订阅拥有 250 个存储帐户。An Azure region supports a maximum of 250 storage accounts per subscription. 达到此限制后,将无法在该区域/订阅组合中创建存储帐户。After this limit is reached, you won't be able to create storage accounts in that region/subscription combination. 顾问会检查你的订阅并提供建议,以便为任何即将达到最大限制的订阅/区域设计更少的存储帐户。Advisor checks your subscriptions and provides recommendations for you to design for fewer storage accounts for any subscription/region that's close to reaching the maximum limit.

考虑增加 VPN 网关 SKU 的大小以解决高 P2S 利用率问题Consider increasing the size of your VPN Gateway SKU to address high P2S use

每个 Azure VPN 网关 SKU 只能支持指定数量的并发 P2S 连接。Each Azure VPN Gateway SKU can support only a specified number of concurrent P2S connections. 如果连接计数接近网关限制,则其他连接尝试可能会失败。If your connection count is close to your gateway limit, additional connection attempts might fail. 如果增加网关的大小,你将能够支持更多并发 P2S 用户。If you increase the size of your gateway, you'll be able to support more concurrent P2S users. 顾问会提供有关增加网关大小的建议和说明。Advisor provides recommendations and instructions for increasing the size of your gateway.

考虑增加 VPN 网关 SKU 的大小以解决高 CPU 利用率问题Consider increasing the size of your VPN Gateway SKU to address high CPU

在高流量负载下,VPN 网关可能会由于高 CPU 利用率而丢弃数据包。Under high traffic load, your VPN gateway might drop packets because of high CPU. 考虑升级 VPN 网关 SKU。Consider upgrading your VPN Gateway SKU. 增加 VPN 网关的大小将确保连接不会因高 CPU 利用率而断开。Increasing the size of your VPN gateway will ensure that connections aren't dropped because of high CPU. 顾问会提供用于主动解决此问题的建议。Advisor provides recommendations to proactively address this problem.

加载时增加批大小,以最大限度提高负载吞吐量、数据压缩和查询性能Increase batch size when loading to maximize load throughput, data compression, and query performance

顾问可检测在向数据库中进行加载时,是否可以通过增加批大小来提高负载性能和吞吐量。Advisor detects whether you can increase load performance and throughput by increasing the batch size when loading into your database. 可以考虑使用 COPY 语句。You could consider using the COPY statement. 如果无法使用 COPY 语句,则在使用 SQLBulkCopy API 或 BCP 等加载实用工具时,考虑增加批大小。If you can't use the COPY statement, consider increasing the batch size when you use loading utilities like the SQLBulkCopy API or BCP. 一条很好的通用规则是,使用在 10 万到 100 万行之间的批大小。A good general rule is to use a batch size that's between 100 thousand and 1 million rows. 增加批大小将提高负载吞吐量、数据压缩和查询性能。Increasing batch size will increase load throughput, data compression, and query performance.

将存储帐户归置到同一区域,以在加载时最大程度地减少延迟Co-locate the storage account in the same region to minimize latency when loading

顾问可检测是否从不同于 SQL 池的区域进行加载。Advisor detects whether you're loading from a region that's different from your SQL pool. 考虑从与 SQL 池位于同一区域内的存储帐户进行加载,以最大程度地减少加载数据时的延迟。Consider loading from a storage account that's in the same region as your SQL pool to minimize latency when loading data. 此更改将有助于最大程度地减少延迟并提高负载性能。This change will help minimize latency and increase load performance.

使用受支持的 Kubernetes 版本Use a supported Kubernetes version

顾问可检测不受支持的 Kubernetes 版本。Advisor detects unsupported versions of Kubernetes.

优化 Azure Database for MySQL、Azure Database for PostgreSQL 和 Azure Database for MariaDB 服务器的性能Optimize the performance of your Azure Database for MySQL, Azure Database for PostgreSQL, and Azure Database for MariaDB servers

解决具有 CPU 瓶颈的 Azure Database for MySQL、Azure Database for PostgreSQL 和 Azure Database for MariaDB 服务器的 CPU 压力问题Fix the CPU pressure of your Azure Database for MySQL, Azure Database for PostgreSQL, and Azure Database for MariaDB servers with CPU bottlenecks

CPU 利用率长时间处于较高状态可能导致工作负荷的查询性能降低。High utilization of the CPU over an extended period can cause slow query performance for your workload. 增加 CPU 大小将有助于优化数据库查询的运行时并提高整体性能。Increasing the CPU size will help to optimize the runtime of the database queries and improve overall performance. 顾问可识别 CPU 利用率较高并可能运行 CPU 受约束工作负荷的服务器,并建议缩放计算。Advisor identifies servers with a high CPU utilization that are likely running CPU-constrained workloads and recommends scaling your compute.

减少 Azure Database for MySQL、Azure Database for PostgreSQL 和 Azure Database for MariaDB 服务器上的内存约束或迁移到内存优化 SKUReduce memory constraints on your Azure Database for MySQL, Azure Database for PostgreSQL, and Azure Database for MariaDB servers, or move to a Memory Optimized SKU

低缓存命中率可能导致查询性能降低和 IOPS 增加。A low cache hit ratio can result in slower query performance and increased IOPS. 这种情况可能是由错误的查询计划或内存密集型工作负荷导致的。This condition could be caused by a bad query plan or a memory-intensive workload. 修复查询计划或 增加 Azure Database for PostgreSQL、Azure Database for MySQL 或 Azure Database for MariaDB 服务器的内存将有助于优化数据库工作负荷的执行。Fixing the query plan or increasing the memory of the Azure Database for PostgreSQL, Azure Database for MySQL, or Azure Database for MariaDB server will help optimize the execution of the database workload. Azure 顾问可识别受此高缓冲池变动影响的服务器。Azure Advisor identifies servers affected by this high buffer pool churn. 建议执行以下操作之一:It recommends that you take one of these actions:

  • 修复查询计划Fix the query plan
  • 迁移到具有更多内存的 SKUMove to an SKU that has more memory
  • 增加存储大小以获得更多 IOPS。Increase storage size to get more IOPS.

使用 Azure Database for MySQL 或 Azure Database for PostgreSQL 只读副本横向扩展读取密集型工作负荷的读取Use an Azure Database for MySQL or Azure Database for PostgreSQL read replica to scale out reads for read-intensive workloads

顾问使用基于工作负荷的启发(如过去 7 天内服务器上的读写比)来识别读取密集型工作负荷。Advisor uses workload-based heuristics like the ratio of reads to writes on the server over the past seven days to identify read-intensive workloads. 具有较高读/写比的 Azure Database for PostgreSQL 或 Azure Database for MySQL 资源可能导致 CPU 或内存争用并导致查询性能降低。An Azure Database for PostgreSQL or Azure Database for MySQL resource with a high read/write ratio can result in CPU or memory contentions and lead to slow query performance. 添加 副本有助于将读取横向扩展到副本服务器,并防止主服务器上的 CPU 或内存限制。Adding a replica will help to scale out reads to the replica server and prevent CPU or memory constraints on the primary server. 顾问可识别具有读取密集型工作负荷的服务器,并建议添加 只读副本 以卸载某些读取工作负荷。Advisor identifies servers with read-intensive workloads and recommends that you add a read replica to offload some of the read workloads.

将 Azure Database for MySQL、Azure Database for PostgreSQL 或 Azure Database for MariaDB 服务器缩放为更高的 SKU,以防止连接约束Scale your Azure Database for MySQL, Azure Database for PostgreSQL, or Azure Database for MariaDB server to a higher SKU to prevent connection constraints

到数据库服务器的每个新连接都会占用内存。Each new connection to your database server occupies memory. 如果由于内存 上限而导致与服务器的连接失败,则数据库服务器的性能会下降。The database server's performance degrades if connections to your server are failing because of an upper limit in memory. Azure 顾问可识别运行时连接失败很多次的服务器。Azure Advisor identifies servers running with many connection failures. 建议通过执行以下操作之一来升级服务器的连接限制,以向服务器提供更多内存:It recommends upgrading your server's connection limits to provide more memory to your server by taking one of these actions:

  • 纵向扩展计算资源。Scale up compute.
  • 使用内存优化 SKU,其中每个核心具有更多的计算资源。Use Memory Optimized SKUs, which have more compute per core.

将缓存缩放到其他大小或 SKU,以提高缓存和应用程序性能Scale your cache to a different size or SKU to improve cache and application performance

如果运行时不存在高内存压力、高服务器负载或高网络带宽,则缓存实例性能最佳。Cache instances perform best when they're not running under high memory pressure, high server load, or high network bandwidth. 这些情况可能导致它们变得无响应、出现数据丢失或变得不可用。These conditions can cause them to become unresponsive, experience data loss, or become unavailable. 顾问可识别处于这些情况下的缓存实例。Advisor identifies cache instances in these conditions. 建议执行以下操作之一:It recommends that you take one of these actions:

  • 应用最佳做法,以降低内存压力、服务器负载或网络带宽。Apply best practices to reduce the memory pressure, server load, or network bandwidth.
  • 缩放到具有更大容量的其他大小或 SKU。Scale to a different size or SKU that has more capacity.

将包含流量的区域添加到 Azure Cosmos DB 帐户Add regions with traffic to your Azure Cosmos DB account

顾问可检测包含来自当前未配置区域的流量的 Azure Cosmos DB 帐户。Advisor detects Azure Cosmos DB accounts that have traffic from a region that isn't currently configured. 建议添加该区域。It recommends adding that region. 这样做可以减少来自该区域的请求的延迟,并在发生区域中断的情况下确保可用性。Doing so improves latency for requests coming from that region and ensures availability in case of region outages. 了解有关如何使用 Azure Cosmos DB 全局分布数据的详细信息。Learn more about global data distribution with Azure Cosmos DB.

使用自定义的已包括或已排除路径配置 Azure Cosmos DB 索引编制策略Configure your Azure Cosmos DB indexing policy by using custom included or excluded paths

顾问可识别使用默认索引编制策略,但可从自定义索引编制策略受益的 Azure Cosmos DB 容器。Advisor identifies Azure Cosmos DB containers that are using the default indexing policy but could benefit from a custom indexing policy. 此判断基于工作负荷模式。This determination is based on the workload pattern. 默认索引编制策略为所有属性编制索引。The default indexing policy indexes all properties. 在查询筛选器中使用显式已包括或已排除路径的自定义索引策略可减少进行索引编制时使用的 RU 和存储。A custom indexing policy with explicit included or excluded paths used in query filters can reduce the RUs and storage consumed for indexing. 了解有关如何修改索引编制策略的详细信息。Learn more about modifying index policies.

将 Azure Cosmos DB 查询页大小 (MaxItemCount) 设置为 -1Set your Azure Cosmos DB query page size (MaxItemCount) to -1

Azure 顾问可识别使用的查询页大小为 100 的 Azure Cosmos DB 容器。Azure Advisor identifies Azure Cosmos DB containers that are using a query page size of 100. 建议使用的页大小为 -1,以提高扫描速度。It recommends using a page size of -1 for faster scans. 详细了解 MaxItemCount。Learn more about MaxItemCount.

考虑在 HBase 群集中使用加速写入功能来提高群集性能Consider using Accelerated Writes feature in your HBase cluster to improve cluster performance

Azure 顾问会分析过去 7 天内的系统日志,并识别群集是否遇到以下情况:Azure Advisor analyses the system logs in the past 7 days and identifies if your cluster has encountered the following scenarios:

  1. 高 WAL 同步时间延迟High WAL sync time latency
  2. 高写入请求计数(1 小时内至少 3 个,超过 1000 个平均写入请求/秒/节点)High write request count (at least 3 one hour windows of over 1000 avg_write_requests/second/node)

这些情况指示你的群集遇到了高写入延迟。These conditions are indicators that your cluster is suffering from high write latencies. 这可能是由于群集上执行繁重工作负荷所致。若要提高群集性能,你可能需要考虑使用 Azure HDInsight HBase 提供的加速写入功能。This could be due to heavy workload performed on your cluster.To improve the performance of your cluster, you may want to consider utilizing the Accelerated Writes feature provided by Azure HDInsight HBase. 适用于 HDInsight Apache HBase 群集的加速写入功能可将高级 SSD 托管磁盘附加到每个 RegionServer(工作器节点),而不是使用云存储。The Accelerated Writes feature for HDInsight Apache HBase clusters attaches premium SSD-managed disks to every RegionServer (worker node) instead of using cloud storage. 因此,请提供应用程序的低写入延迟和更好的复原能力。As a result, provides low write-latency and better resiliency for your applications. 若要详细了解此功能,请了解详细信息To read more on this feature, learn more

查看 Azure 数据资源管理器表缓存周期(策略)以获得更好的性能(预览版)Review Azure Data Explorer table cache-period (policy) for better performance (Preview)

此建议显示 Azure 数据资源管理器表,这些表具有大量查询,可追溯到配置的缓存周期(策略)之外(你将看到按访问缓存外数据的查询百分比排序的前 10 个表)。This recommendation surfaces Azure Data Explorer tables which have a high number of queries that look back beyond the configured cache period (policy) (You will see the top 10 tables by query percentage that access out-of-cache data). 提高群集性能的建议操作:将此表上的查询限制为所需的最小时间范围(在定义的策略内)。The recommended action to improve the cluster's performance: Limit queries on this table to the minimal necessary time range (within the defined policy). 或者,如果需要整个时间范围内的数据,请将缓存周期增加为建议值。Alternatively, if data from the entire time range is required, increase the cache period to the recommended value.

通过优化 MySQL 临时表的大小来提高性能Improve performance by optimizing MySQL temporary-table sizing

顾问分析指示,由于临时表参数设置较低,MySQL 服务器可能会产生不必要的 I/O 开销。Advisor analysis indicates that your MySQL server may be incurring unnecessary I/O overhead due to low temporary-table parameter settings. 这可能会导致不必要的基于磁盘的事务并降低性能。This may result in unnecessary disk-based transactions and reduced performance. 建议增大“tmp_table_size”和“max_heap_table_size”参数值,以减少基于磁盘的事务数。We recommend that you increase the 'tmp_table_size' and 'max_heap_table_size' parameter values to reduce the number of disk-based transactions. 了解详细信息Learn more

在服务器组中分布数据,以在节点间分布工作负载Distribute data in server group to distribute workload among nodes

顾问可识别数据未进行分布,而是仍保留在协调器中的服务器组。Advisor identifies the server groups where the data has not been distributed but stays on the coordinator. 基于这一点,顾问建议在服务器组中的工作器节点上分布数据,以获得完整的超大规模 (Citus) 权益。Based on this, Advisor recommends that for full Hyperscale (Citus) benefits distribute data on worker nodes for your server groups. 这将通过利用服务器组中每个节点的资源来提高查询性能。This will improve query performance by utilizing resource of each node in the server group. 了解详细信息Learn more

如何访问顾问中的性能建议How to access performance recommendations in Advisor

  1. 登录 Azure 门户,并打开顾问Sign in to the Azure portal, and then open Advisor.

  2. 在顾问仪表板中,选择“性能”选项卡。On the Advisor dashboard, select the Performance tab.

后续步骤Next steps

若要了解有关顾问建议的详细信息,请参阅以下资源:To learn more about Advisor recommendations, see: