[[["容易理解","easyToUnderstand","thumb-up"],["確實解決了我的問題","solvedMyProblem","thumb-up"],["其他","otherUp","thumb-up"]],[["難以理解","hardToUnderstand","thumb-down"],["資訊或程式碼範例有誤","incorrectInformationOrSampleCode","thumb-down"],["缺少我需要的資訊/範例","missingTheInformationSamplesINeed","thumb-down"],["翻譯問題","translationIssue","thumb-down"],["其他","otherDown","thumb-down"]],["上次更新時間：2025-08-18 (世界標準時間)。"],[[["\u003cp\u003eThe Dataflow monitoring dashboard provides a project-level overview of all Dataflow jobs, using charts to display metrics and help identify issues like quota errors, autoscaling anomalies, and slow streaming jobs.\u003c/p\u003e\n"],["\u003cp\u003eThe dashboard uses Cloud Monitoring to access job metrics, and users can customize the dashboard to filter jobs, edit widgets, and select metrics displayed in the charts.\u003c/p\u003e\n"],["\u003cp\u003eUsers must have the Monitoring Viewer IAM role, which includes the \u003ccode\u003emonitoring.timeSeries.list\u003c/code\u003e permission, to view the graph data in the dashboard.\u003c/p\u003e\n"],["\u003cp\u003eThe dashboard features several default charts, including running jobs, workers per job, quota exceeded errors, CPUs per job, system latency, data freshness, and max backlog bytes, which support troubleshooting for both batch and streaming jobs.\u003c/p\u003e\n"],["\u003cp\u003eIf the dashboard displays a "No data is available" message, users should adjust the time range selector; to restore default settings, users can select "Predefined" from the dashboard menu after customizing.\u003c/p\u003e\n"]]],[],null,["The Dataflow web-based monitoring interface includes a dashboard\nthat monitors your Dataflow jobs at the project level. The charts\nshow data for all of the jobs in one project.\n\n[Go to dashboard](https://console.cloud.google.com/dataflow/monitoring)\n\nThe dashboard can help you with the following tasks:\n\n- Detect and identify the source of quota errors.\n- Detect anomalous horizontal autoscaling in a job.\n- Identify slow or stuck streaming jobs.\n\nThe dashboard uses\n[Cloud Monitoring](/dataflow/docs/guides/using-cloud-monitoring) to access\nDataflow job metrics. To customize the information displayed in\nthe charts, use [Metrics Explorer](/monitoring/charts/metrics-selector).\n\nFeatures\n\nThe dashboard includes the following features:\n\n- Choose which jobs appear in the dashboard by using regular expressions.\n- Access the job details page from individual charts.\n- Customize the dashboard widgets and charts.\n\nRequired roles\n\n\nTo get the permission that\nyou need to see the graph data,\n\nask your administrator to grant you the\n\n\n[Monitoring Viewer](/iam/docs/roles-permissions/monitoring#monitoring.viewer) (`roles/monitoring.viewer`)\nIAM role.\n\n\nFor more information about granting roles, see [Manage access to projects, folders, and organizations](/iam/docs/granting-changing-revoking-access).\n\n\nThis predefined role contains the\n` monitoring.timeSeries.list`\npermission,\nwhich is required to\nsee the graph data.\n\n\nYou might also be able to get\nthis permission\nwith [custom roles](/iam/docs/creating-custom-roles) or\nother [predefined roles](/iam/docs/roles-overview#predefined).\n\nAccess the dashboard\n\nTo access the dashboard, follow these steps:\n\n1. [Sign in](https://console.cloud.google.com/) to the Google Cloud console.\n2. Select your Google Cloud project.\n3. Open the navigation menu.\n4. In **Analytics** , click **Dataflow**.\n5. In the Dataflow navigation menu, click **Monitoring**.\n\n [Go to dashboard](https://console.cloud.google.com/dataflow/monitoring)\n\nDashboard metrics\n\nBy default, the following time-series charts appear in the dashboard. For more\ninformation about the metrics displayed, see\n[Job metrics](/dataflow/docs/guides/using-monitoring-intf).\n\nThe following charts apply to batch and streaming jobs:\n\n- **Running jobs**. Shows the number of active jobs running in the project. This chart indicates the overall Dataflow activity in the project over time.\n- **Workers per job (top 25)**. Shows the current worker counts for the 25 most parallelized jobs. This chart is useful for understanding resource allocation and identifying high-workload jobs. You can also see if jobs have unexpected scaling behavior.\n- **Total count of vCPUs**. Shows the total number of virtual CPUs (vCPUs) in use across all jobs in the project. The total number of vCPUs affects Compute Engine quotas.\n- **vCPUs per job (Top 25)**. Show the 25 jobs that consume the most vCPU resources. This chart highlights potentially expensive jobs.\n- **Total count of vCPUs**. Shows a project-wide aggregate of vCPUs in use. This chart gives a high-level view of the Compute Engine resources that your jobs consume.\n- **Quota exceeded errors** . Reports any instances where [Dataflow quotas](/dataflow/quotas) or [Compute Engine quotas](/compute/quotas-limits) have been reached. This chart can help you to find potential job failures or scaling slowdowns.\n\nThe following charts apply to streaming jobs:\n\n- **Average system latency** . Shows the average [system latency](/dataflow/docs/guides/using-monitoring-intf#system_latency_streaming), which reflects the typical delay experienced by data as it passes through source stages. This chart can indicate potential input bottlenecks. Use this chart to identify streaming jobs that have an unusual delay between when data appears in a source and when the data is written to all sinks.\n- **Top 25 jobs by system lag**. Shows the 25 streaming pipelines with the highest system lag, which is the longest amount of time that data spends being processed or awaiting processing. This chart can indicate potential real-time processing bottlenecks.\n- **Top 25 jobs by data watermark lag per stage (freshness)** . Shows the 25 streaming jobs with the largest watermark lag. The *watermark lag* for a stage is the difference between the latest event time received by the stage and the watermark. This chart can indicate potential bottlenecks at per-stage granularity. Use this chart to find streaming jobs that might be slow or stuck. For more information, see [Troubleshoot slow or stuck streaming-jobs](/dataflow/docs/guides/troubleshoot-slow-streaming-jobs).\n- **Top 25 jobs by SECU usage** . Shows the 25 streaming jobs that consume the most [Streaming Engine Compute Units](/dataflow/pricing#streaming-compute-units). Use this chart to measure the cost and intensity of your streaming jobs that use [resource-based billing](/dataflow/docs/streaming-engine#compute-unit-pricing).\n- **Top 25 jobs by user processing latencies (per stage)**. Shows the 25 streaming jobs where user-defined code in processing stages takes the longest. Use this chart to find potential performance bottlenecks in your application logic.\n- **Max backlog bytes (top 25)**. Shows the 25 streaming jobs with the largest volume of unprocessed data waiting at any stage. This chart can indicate potential input overload or slow processing.\n\nFor more information about working with charts, see\n[Explore charted data](/monitoring/charts/working-with-charts).\n\nCustomize the dashboard\n\nYou can customize the dashboard contents and the information displayed in the\ncharts. When you edit the dashboard, a new, customized dashboard is created.\n\nThe dashboard uses Cloud Monitoring to access Dataflow job\nmetrics. Use the Cloud Monitoring tools to customize the charts.\n\n1. Open the dashboard and click **Customize Dashboard**.\n2. Modify your dashboard.\n - To filter the jobs that display on the dashboard, see [Add temporary filters to a custom dashboard](/monitoring/charts/filter-dashboard) and [Add permanent filters to a custom dashboard](/monitoring/dashboards/filter-permanent).\n - To edit or remove widgets, see [Manage dashboard widgets](/monitoring/charts/manage-widgets).\n - To edit the contents of the charts, see [Select metrics for charts on dashboards](/monitoring/charts/metrics-selector).\n - To add charts to the dashboard, see [Add charts and tables to a custom dashboard](/monitoring/charts).\n3. Click **Save** , and then click **View customized dashboard**.\n\nAfter you create a customized dashboard, to return to the default dashboard,\nin the **Dashboard** menu, select **Predefined**.\n\nFor an example of adding a custom metrics chart to the dashboard, see\n[Customize the Dataflow monitoring dashboard](/dataflow/docs/guides/customize-monitoring-dashboard).\n\nTroubleshooting\n\nThis section provides instructions for troubleshooting common issues\n\nNo data is available\n\nWhen you open your dashboard, one or more charts shows the following message: \n\n No data is available for the selected time frame.\n\nThis message appears when the time period covered in the charts doesn't have any data. To resolve this issue, change or expand the time range.\n\nTo change the displayed time range, on the chart, click **Explore data**, and then use the time-range selector.\n\nUnable to restore deleted widgets\n\nWhen you remove a widget from the dashboard, you create a customized dashboard.\nAfter you create a customized dashboard, to return to the default dashboard,\nin the **Dashboard** menu, select **Predefined**.\n\nUnable to view charts\n\nTo see the graph data, you need the `monitoring.timeSeries.list`\npermission. For more information, see [Required roles](#requirements).\n\nWhat's next\n\n- Learn more about individual [job metrics](/dataflow/docs/guides/using-monitoring-intf).\n- Explore metrics with [Cloud Monitoring](/dataflow/docs/guides/using-cloud-monitoring).\n- Learn how to [troubleshoot Dataflow pipelines](/dataflow/docs/guides/troubleshooting-your-pipeline)."]]
