Resolution
Use the below options to workaround the issue:
Option 1
Reduce the complexity and size of the flow:
- Remove all of the columns you don't need in each input node.
- Shrink the field data you don't plan on using (filter using a calculation, selected value, wildcard match, null values).
- Avoid Custom SQL as it bypasses our query processor which optimizes for performance.
- If joining static flat files with databases, move the flat file data into the database and do the join there instead of in Prep.
- Move filters, aggregates before the join to shrink the table and help performance during the join step.
- Add aggregate steps when possible to reduce multiple values down to a single value.
Option 2
Split the flow up into multiple smaller flows and use the outputs as inputs for a final flow. Utilize linked tasks to run the final flow after the initial flows have completed. See Schedule linked tasks for more information.
Option 3
Run the flow manually in Prep Builder and then republish the flow to Tableau Cloud.
Option 4
Automate refreshes using the Prep Command Line. See Refresh flow output files from the command line for more information.