Tableau Prep Flow Authoring
The Tableau Prep Flow Authoring process provides the interactive Prep Flow experience in the browser, which lets users create and interact with flows on Tableau Server to clean and prepare data. For more information, see Create and Interact with Flows on the Web. It was added in version 2020.4.
|Status||Status of the |
|Logging||Logs generated by the |
Tableau Prep Flow Authoring is enabled by default. It uses these components:
- Tableau Prep Minerva Service: Used by Tableau Prep Flow Authoring for querying data sources. By default, it is automatically enabled on a node where Tableau Prep Flow Authoring is enabled.
- Data Engine: The Data Engine creates data extracts and processes queries. By default, it is automatically enabled on a node where Tableau Prep Flow Authoring is enabled.
- Gateway: The Gateway process handles all requests to Tableau Server from browsers, Tableau Desktop and other clients. By default, it is automatically enabled on a node where Tableau Prep Flow Authoring is enabled.
Performance and Scale Recommendations
Tableau Prep Flow Authoring can be CPU and RAM intensive. It is recommended to monitor the server and adjust your deployment as necessary by isolating flow authoring to a separate node or adding server resources.
Isolate flow authoring to a separate node
Running Tableau Prep Flow Authoring on a separate node will isolate flow authoring workflows from other Tableau workloads. If you have a busy server with many extract refreshes and a lot of viz editing and viewing, and you don't want to cause disruption to this workload, then it is recommended to isolate Prep Flow Authoring to its own server node. This means that all flow editing is directed to a dedicated node. For more information see Configure Nodes.
Note: Prep Web Authoring uses shared services such as the Application Server (VizPortal), the PostgreSQL repository, the Cache Server and Hyper. If those services are already at capacity, they might also need additional resources because of the additional Prep Web Authoring load.
If you are running Tableau Server on a multi-node cluster, you can dedicate one or more nodes for running Backgrounder. Using the Backgrounder process, you can isolate background workloads such as Tableau Prep Conductor from all your interactive workloads such as Prep Flow Authoring and VizQL Server. For more information, see Tableau Server Backgrounder Process and Workload Management through Node Roles.
For core and user-based deployments, it’s strongly recommended that at least one node be dedicated to flows for best performance. While you can run flows on any licensed Server core, the additional resource cores purchases should only run Tableau Prep Conductor, and not extract refreshes or VizQL processes.
As you need more resources, you can add more nodes to your server environment. When planning your nodes, several factors can influence your decision on how much additional hardware you need to allocate.
The main things to consider when planning your nodes are:
- The number of concurrent users or concurrent sessions you expect during peak hours. The number of concurrent sessions per user can be set using the TSM option
maestro.sessionmanagement.maxConcurrentSessionPerUser. For more information, see tsm configuration set Options.
- The number of data input nodes that your flows have on average and the amount of data they have. The maximum row sampling limit can be set using the TSM option
maestro.app_settings.sampling_max_row_limit. For more information, see tsm configuration set Options.
- The complexity of the flows being authored and the number of nodes. Node types like joins, unions, aggregates and pivots will, in general, require more resources.
Licence additional offerings
Additional licences for Data Management and Advanced Management are required for configuring nodes.
|OFFERING||Allows you to:|
Configure a node to run only flows, or configure a node to run all jobs except flows. Tableau Prep Conductor must be running on the node.
The Data Management licence includes Tableau Prep Conductor, which enables you to schedule and track flows. The licence is for a single Tableau Server deployment, which can be role-based or core-based.
As a Creator, Data Management is not required to create and edit flows directly on your server.
|Advanced Management||Configure where the different types of workloads are processed through node rules. For example, you can run flows on one node and subscriptions and alerts on another node.|
Topology and Configuration
To make Tableau Prep Flow Authoring highly available, configure two instances of Tableau Prep Flow Authoring on nodes when enabling it.
Here is an example of a two node configuration:
Here is an example of a status page for a two node configuration:
Here is an example of a four node configuration:
Here is an example of a status page for a four node configuration: