Application Server:
When a user opens a view in a client device, that user starts a session on Tableau Server
VizQL Server:
VizQL process sends queries to the data
source, returning a result set which is rendered as images
and presented to the user.
Each VizQL Server has its
own cache that can be shared across multiple users.
Architecture - contd.
Data Server:
Maintains metadata from Tableau Desktop, such as calculations, definitions, and groups
Published data source can be base on
A live connection (cubes are not supported as live connections
A Tableau Data Engine extract
Gateway/ Load Balancer
Routes requests to other components.
Requests that come in from
the client first hit an external load balancer, if one is configured, or the gateway
and are routed to the appropriate process
Data Layer Connection Options - Live and Extract
Data Layer Connection Options - Live
Connection: Live:
Use High Performance capability of the source data store and to get near real-time update
Sends dynamic SQL or Multidimensional Expressions (MDX) to backend data store
This can result in slow experience depending on the backend data store capability
Data Layer Connection Options - Extract
Connection: Extract:
Extracts the data into Tableau Data Engine (Memory and Disk)
Takes the Query Load out of the critical backend data store
Selective extract will help the performance
Make sure you have enought diskspace to avoid:
Tableau Data Engine Error 4: There is not enough space on the disk.
Data Engine Disk Space requirements
As much as the square of the final size of the extract: extract_size ** 2
Example: 32GB extract may require up to 1024GB (1TB) of disk space to create the extract.
