Video | Tableau | Data prep | Tool strategy

Data freshness policies in Tableau Server & Online : New in Tableau 2021.3

Tableau 2021.3 finally lets you dictate exactly how fresh your live-connection data needs to be.

Part ofWhat's New in Tableau 2021.3Tableau Server
Watch on YouTube
  • Data freshness policies only apply to live connections; embedded extracts don't expose the option.
  • Access the setting via the eye icon on a workbook's details, not the three-dot menu.
  • The site default is 12 hours, but you can set freshness from minutes up to weeks.
  • You can schedule a specific time and time zone for data to refresh, similar to extract schedules.
  • Caching happens at multiple points (Tableau and the database), so a 'live' request still may return cached data from the database itself.
  • You can't set multiple refresh times per day; use a recurring interval like every three hours instead.

Tableau 2021.3 lets you set a data freshness policy on a workbook so you can control exactly how often live-connected data is refreshed, giving you precision that was previously hard to get with server/database caching.

Works only on workbooks using a live connection to a data source; extracted workbooks don't expose this setting at all. Demonstrated on Tableau Server/Online using a workbook connected live to Snowflake.

The Breakdown
  1. Confirm you're using a live connection 0:11

    Check your workbook's data sources first: if it's an embedded extract, the freshness policy option won't appear because caching rules don't apply to extracts the same way. This is a hard prerequisite before anything else here works.

  2. Extracted workbooks won't show the option 0:49

    On an extract-based workbook, opening the details view confirms there's no data freshness setting present at all, since the workbook isn't live-connected.

  3. Open the setting via the eye icon, not the three dots 1:23

    The three-dot menu only gives a basic dropdown; you need to click the small eye icon on the workbook to open its details page, which is where the data freshness policy actually lives.

  4. Understand the site default and caching layers 1:45

    If you haven't set anything, the workbook inherits a site default (12 hours in this case). Caching can happen both in Tableau and at the database itself, so even a 'live' setting doesn't guarantee the database isn't serving cached results — it gives certainty about what Tableau is doing, not the whole chain.

  5. Choose live, an interval, or a scheduled time 2:43

    You can set the policy to always live (Tableau queries the database on every request), to a recurring interval anywhere from minutes to weeks, or to a specific scheduled refresh time — pick whichever matches how often the underlying data actually changes.

  6. Schedule refresh by time of day and time zone 3:42

    You can specify particular days and a time (in a chosen time zone) after which data becomes fresh again, similar to setting an extract refresh schedule — useful for aligning freshness with when people actually check the dashboard.

  7. Work around the single-time-per-day limit 4:41

    You can't set multiple specific refresh times in one day; if you need that, use a recurring interval (e.g. every three hours) instead, which covers most real-world needs.

Worth Knowing
  • A 'live' setting only guarantees Tableau sends a fresh request to the database — the database itself may still return cached data, so live doesn't always mean truly live.
  • You can only set one scheduled time per day; there's no way to add multiple discrete refresh times without switching to a recurring interval.
  • The freshness policy input for intervals is a toggle, not free typing, which can be fiddly when picking an exact value.
Use It When

Reach for this when you have a live-connected workbook and need explicit control over how current the data is — for example ensuring it's fully live for critical decisions, or deliberately relaxing freshness to reduce load when the source barely changes.

How this Rollup was made provenance & method

A Rollup is drafted by AI from the video's transcript, then reviewed and edited by Tim. Everything used to produce this one is listed below — the model, the exact prompt, and the source video — so the process is transparent and reproducible.

Transcription
On-device — NVIDIA Parakeet v3 for recent videos, OpenAI Whisper large-v3 for earlier ones. The transcript never leaves the machine or gets published.
Drafting
Claude Sonnet 5 in the cloud, from that transcript.
Prompt
The exact Rollup prompt (v2) — the full system prompt, unedited.
Source video
Watch on YouTube
Drafted
5 July 2026 at 09:38
Reviewed & edited
5 July 2026 at 09:41 · by Tim Ngwena

Model + prompt + video is everything you'd need to recreate a Rollup like this yourself. The one thing we don't share is the transcript.

Rights. The video and its transcript are the property of TN Media Ltd. Unauthorised use or download is prohibited. © TN Media Ltd.