Tableau Cloud Help

Your Tableau Cloud site comes with capacity to support all your users’ analytic needs. A site’s capacity includes capacity for storage and tasks that need to be performed on the site for extracts, metrics, subscriptions and flows.

The capacity information discussed in this topic is specific to each site in your Tableau Cloud Manager (TCM) tenant, the top-level administrative layer that manages your organisation's Tableau Cloud deployment. Depending on your licence type, TCM includes up to 3 sites by default. The number of sites increases with different licence types – up to 10 sites with an Enterprise licence and a maximum of 50 with a Tableau+ licence. For more information about TCM, see Use Tableau Cloud Manager.

Licensing capacity is managed at the tenant level and can be allocated to sites by cloud administrators using site role limits. For more information, see Manage Site Role Limits.

Summary

The following table summarises the capacity allowances for your site by feature. You can find more capacity information by the features listed below.

*Note: To ensure that system resources are equitably distributed across Tableau Cloud, each site has an upper limit on the number of jobs that can run at any point in time. However, this limit is not a guarantee of capacity. Tableau Cloud allocates system resources to optimise jobs, and resources are adjusted to start a job within 15 minutes or less from the scheduled start time.

Note: The behaviour of the capacity limits specific to each site within Tableau Cloud Manager is subject to change as we evolve our infrastructure.

Storage capacity

A Tableau Cloud site comes with site and individual content storage capacities.

• Site storage: A site comes with 1 TB of storage capacity. Workbooks, published data sources and flows count toward this storage capacity.

• Individual workbook, published data source and flow size: An individual workbook, data source (live or extract) or flow published to your site can have a maximum size of 15 GB.

  Note: If your extract data source exceeds 10 GB in size, we recommend that you consider either using live connection to the database or aggregate the data in the extract to reduce its size. Frequently republishing or refreshing large extracts can be time intensive and usually indicates that more efficient data freshness strategies should be considered.

Visualisation capacity

Each site comes with designated capacity to load and query individual visualisations, also known as views. You can find more information about each capacity type below.

• Load time: A site has capacity to spend up to 5 minutes to load a view.

• Idle timeout: If the user hasn’t interacted with or edited a view for 120 consecutive minutes, the user must reload (for example, refresh the page or click the Reconnect button if a banner displays) the view or saved draft to interact with it again.

• User request rate: Each user on a site has the capacity to make up to 600 requests per hour to load and refresh views. When this capacity has been reached, the user can't interact with, edit and save the view. After an hour, the view can be refreshed and new views can be loaded.

  Note: If the cumulative number of user requests exceeds the site-level request rate, users can’t interact with, edit or save views.

• Query usage size: A site has capacity to use no more than 20 GB of memory to query a view that uses an extract data source. The amount of memory needed to query a view can vary depending on the complexity of the data manipulations that Tableau needs to perform to generate the marks in the view.

  Note: Only queries against extract data sources count toward query usage capacity. Queries against data sources that use live connections to the underlying data do not. However, queries against data sources that use live connections might be subject to other query usage capacity enforced by underlying databases.

Views that exceed these capacity types might be cancelled to ensure enough system resources are available so that other views on the site can load.

Tips for optimising visualisation capacity

Views that exceed visualisation capacity can indicate the queries that Tableau must run in order to generate the view might be too complex. In such cases, you can use the following tips to help optimise queries thereby minimising the time it takes to load views on Tableau Cloud. Most tips require that you edit the data source or workbook and republish to Tableau Cloud for changes to take effect.

For more tips, see the Workbook Performance(Link opens in a new window) topic in the Tableau Help.

• Use data source filters: When you add a filter on a data source, you reduce the amount of data that needs to be generated in the view. Reducing the amount of data can help minimise load times of views. For more information, see Filter Data from Data Sources(Link opens in a new window) in the Tableau Help.

• Use context filters: Similar to data source filters, context filters reduce the data in the data source but at the sheet level. Context filters work independently of other filters and can significantly reduce the amount of data that needs to be generated in the view. Reducing the amount of data can help minimise load times of views. For more information, see Improve View Performance with Context Filters in the Tableau Help.

• Decrease date ranges: As a rule, dates require more computing power than other types of data like numbers and Booleans. If the view contains date filters, you can reduce the date ranges or the number of dates that need to be generated to help minimise the load times of views.

• Reduce marks: Because Tableau must calculate the marks in order to generate the view, the number of marks can impact how quickly the view can load. There are several ways you can reduce marks, including using some of the filters listed above or using sets to filter, custom SQL to aggregate data, and more. For more information, see Reduce the number of marks in the view(Link opens in a new window).

View Acceleration capacity

Each Tableau site has the following View Acceleration capacity:

• Accelerated views: The default number of accelerated views per site is 30. For each additional Creator Licence, 20 more views can be accelerated. The maximum number of accelerated views is 750.

• Accelerated refreshes: Accelerated views are regenerated in alignment with the following four data update events:

  • Data freshness policy: The acceleration timestamp is out of sync with the data freshness policy.
  • Extract refreshes: An extract refresh completes, signalling fresher data.
  • Accelerated view update: An accelerated view is updated by the owner.
  • View opted-in: A view is originally opted in for acceleration.

  Each unique event triggers a pre-computation job to be queued. Backgrounder jobs then pick up and complete the pre-computation and then store query results as a materialised view. Views that belong to the same workbook are refreshed in one job. To limit resource consumption, the maximum number of pre-computation jobs that can be run is limited to 12 jobs per day. For more information, see Set a Data Freshness Policy.
  

• Acceleration memory usage size: A site has capacity to use no more than 20 GB of memory for pre-computing and fetching the workbook's data in a backgrounder job. The amount of memory needed varies depending on the complexity of the query(ies) that are associated with the view that is being accelerated.

Job runtime capacity

Job runtime capacity is the period of time that a job, or an instance of a task type, is allowed to run before it's cancelled to help protect available system resources. Each site comes with job runtime capacities for the following job types:

• Extract refreshes
• Subscriptions
• Flows
• Metrics
• View Acceleration

Each extract refresh or metrics job type has a maximum runtime of two hours (120 minutes or 7,200 seconds). Subscription or view Acceleration has a maximum runtime of 30 minutes (1,800 seconds).

Note about flows: For sites that don’t have Data Management, the maximum runtime for flow jobs for a site is 60 minutes multiplied by the number of creator licences up to a maximum of 24 hours. If your site has large flows workloads, consider purchasing Data Managementto get more capacity. For sites with Data Management, you can run flows for up to 24 hours per day per Resource Block. If you exceed the time limit, no new jobs can be started until the next day. Additional capacity for concurrent flow jobs can be purchased. For more information, see Resource Blocks.

Note about extract refreshes: If a refresh job reaches its maximum runtime, you see a timeout error. For more information about the error and ways you can modify extracts to keep refresh jobs within the runtime capacity, see Time limit for extract refreshes.

Concurrent jobs capacity

Concurrent jobs capacity is the number of jobs, or instances of a task type that can run at the same time. Each site comes with concurrent capacities for the following job types:

• Extract refreshes
• Subscriptions
• Flows
• Metrics

You can find more information about the concurrent jobs capacity for each job type below.

• Extract refreshes: A site has an upper limit of 10 extracts that it can refresh at any given point in time. Depending on available system resources, the wait time (i.e. the time before a job starts) is adjusted to start a job within 15 minutes or less from the scheduled start time. Jobs that count towards refresh capacity include scheduled refreshes, manual refreshes, extract creation and command line or API calls that trigger refreshes, including appending data incrementally.

• Subscriptions: A site has an upper limit of 10 subscriptions that it can run at any given point in time. Depending on available system resources, the wait time (i.e. the time before a job starts) is adjusted to start a job within 15 minutes or less from the scheduled start time. Jobs that count towards subscriptions capacity include generating emails from subscriptions created through Tableau Cloud directly or API calls that generate the email subscriptions.

• Flows: If licensed with Data Management, a site has a default capacity of 1 concurrent flow job per Resource Block. Resource Blocks support flow jobs. If not licensed with Data Management, a site has a default capacity of 1 concurrent flow job, which is the equivalent of having one Resource Block. If there are no Resource Blocks available when a flow job starts, the job goes into a queue to wait for a Resource Block to be available. A flow job might also go into a queue during busy hours even when there are available Resource Blocks to ensure that system resources remain available to support flow jobs across all of Tableau Cloud. Additional capacity for concurrent flow jobs can be purchased. For more information, see Resource Blocks. Each user can have a maximum of 4 active flow web editing sessions at one time. If more sessions are opened, the oldest session will be closed.

• Metrics: Each site has the capacity to refresh up to 10 metrics concurrently. For more information about how metrics refresh, see Create and Troubleshoot Metrics (Retired).

Note: If a site exhausts its concurrent capacity for a job, other jobs that are in the queue remain in a pending state until one or more jobs have completed.

Daily jobs capacity

Daily jobs capacity is the number of shared hours each site can spend per Creator licence a day to perform all jobs on the site. A server process, called backgrounder, initiates and performs these jobs. A site with more Creator licences gets more daily jobs (backgrounder) capacity to meet the needs of a larger site population.

Each site comes with daily jobs (backgrounder) capacity to perform both extract refresh and metric refresh jobs.

For example:

Daily jobs capacity resets each day at 24:00:00 UTC (coordinated universal time).

About extract refreshes

Refresh jobs that count towards daily jobs capacity include full and incremental refreshes and extract creation, which can be initiated by scheduled refreshes, manual refreshes and certain command line or API calls.

Notes:

• As long as your site has time remaining in its daily refresh capacity, a refresh job will start.
• Site admins receive email notifications when a site reaches 70%, 90% and 100% of its daily refresh capacity.
• If a site exhausts its daily refresh capacity, any future extract refreshes are cancelled and email notifications are sent to content owners alerting them of the refresh cancellation.

Tips for optimising extract refreshes capacity

If you're managing an extract-heavy environment, Tableau recommends following some best practices to make the most efficient use of your site capacity.

• Stop refreshing unused extracts. One of the best ways to reclaim capacity for your site is to stop automatic refreshes, either through schedules on Tableau Cloud or through scripts of extracts that aren’t being used. For more information about deleting a refresh schedule, see Manage Refresh Tasks.

• Reduce the frequency of refreshes. Another method of reclaiming capacity for your site is to reduce the frequency of extract refreshes. For example, instead of refreshing an extract hourly, consider refreshing an extract daily or only during business hours when fresh data is most useful. For more information about changing a refresh schedule, see Manage Refresh Tasks.

• Use incremental refreshes instead of full refreshes. To decrease the amount of time an extract takes to refresh, consider performing an incremental refresh of the extract instead of a full refresh. For more information, see Schedule Refreshes on Tableau Cloud.

  Note: Changing from a full to an incremental refresh can be done from Tableau Cloud only if the extract was configured for incremental refresh in Tableau Desktop before the extract was published. For more information about incremental refreshes, see Configure an incremental extract refresh(Link opens in a new window) in the Tableau User Help.

• Use live connections instead of extract connections. For views, dashboards and workbooks whose data needs to be updated frequently, consider setting up the data source to use a live connection instead of an extract connection.

• Optimise the data in the extract. Improving the performance of an extract can also help reduce the amount of time an extract takes to refresh. There are a number of changes you can make to the extract's data to help its overall performance, such as removing unused fields, using filters to remove unused rows, changing date ranges, etc.

About metric refreshes

Metrics that rely on extract data sources refresh when their extract counterparts refresh and therefore count towards daily job capacity.

Note: Metrics that rely on live data sources refresh every 60 minutes and don't count towards daily job capacity.

Jobs initiated by command line and API calls

In addition to the capacity described above, a site comes with designated capacity for command line and API calls.

• Initiate extract refresh jobs: A site comes with designated capacity for tabcmd (refreshextracts(Link opens in a new window)), REST API (Run Extract Refresh Task(Link opens in a new window)), and Tableau Client (Python) library(Link opens in a new window) calls to initiate extract refresh jobs. A site allows 20 total calls per one-hour period.

• Query or cancel jobs: A site comes with a designated capacity for REST API (Query Job(Link opens in a new window) and Cancel Job(Link opens in a new window)), and Tableau Client (Python) library(Link opens in a new window) calls that query status information for asynchronous jobs, or cancel pending and in-progress jobs. A site allows 20000 calls per 24-hour period.

• Initiate publish, update or refresh jobs: A site comes with designated capacity for tabcmd (publish(Link opens in a new window)), REST API (Publish Workbook(Link opens in a new window), Publish Data Source(Link opens in a new window), Update Workbook(Link opens in a new window), Update Data Source(Link opens in a new window), Update Data Source Now(Link opens in a new window)), and Tableau Client (Python) library(Link opens in a new window) calls that initiate publish, update or refresh jobs for workbooks and data sources. A site allows 4000 calls per 24-hour period.

Note: If the number of calls exceed the command line or API calls capacity described above, you see one of the following errors: Unknown 429 or ApiCallError: 429000: Too Many Requests – Too many requests for <job type> ‘api.rest.refresh_extracts’. Please retry after 146 second(s).

Monitor capacity

There are a few tools you can use to help you monitor your site’s capacity: Jobs page and Admin views.

Jobs page

The Jobs page gives you detail about the unique instances of backgrounder tasks, called jobs, within the past 24 hours. You can use the jobs page to monitor extract refreshes, subscriptions and flow jobs.

1. Sign in to Tableau Cloud using your site admin credentials.

2. In the left navigation pane, click Jobs.

3. From the filter drop-down menu, under Task Type, select the job type you want to monitor.

For more information on the Jobs page, see Managing Background Jobs in Tableau Cloud topic.

Admin views

Admin views help you monitor different types activity on your site. You can find more capacity usage information, using admin views, for the features listed below.

• Storage: You can monitor your site’s storage capacity and check which workbooks, data sources and flows take up the most space on your site using the Admin Insights “Stats for Space Usage” dashboard. Select Explore > Admin Insights > Admin Insights Starter > Stats for Space Usage.

  For more information about Admin Insights, see Use Admin Insights to Create Custom Views.

• Flows: You can monitor the performance of flow runs. Select Site Status > Flow Performance History. Using the dashboard, under Flow History, click on a mark to see details about the flows job.

  For more information about this admin view, see Administrative Views for Flows.

• Metrics: You can monitor your site’s metrics jobs by using the ‘Background Tasks for Non Extracts’ admin view. Select Site Status > Background Tasks for Non Extracts and then, from the Tasks filter drop-down, select Update all metrics on a view tick box.

  For more information about this admin view, see Background Tasks for Non Extracts.