Four lines every morning on which dashboards are current
Which extracts refreshed, which one last refreshed 61 hours ago, and what that means for the ten o’clock meeting. Waiting before anyone quotes a number.
A Tableau extract fails to refresh. The workbook still renders. The numbers are Thursday’s. Somebody presents them on Monday and the meeting proceeds normally, because a chart with old data looks exactly like a chart with new data.
A companion checks freshness before anybody presents and turns that into four lines: three workbooks current, one last refreshed 61 hours ago. It works through an account you connect and checks every workbook on the site, not the three you remembered.
What is the freshness check doing?
Reading when each extract last refreshed and comparing that to when it was supposed to.
Two facts, neither of them visible to the person looking at the chart. "These three workbooks are current, this one last refreshed 61 hours ago" gives a meeting the caveat it needs.
Run every morning by a routine, that is four lines. Run by a person it is a login, a project browse and a decent amount of goodwill, which is why it happens the week after an incident and then stops. Recurring reads like this are ordinary operations hygiene.
What else is worth handing over?
The inventory.
Any Tableau site older than two years has workbooks nobody opens, data sources duplicated three ways because someone needed one extra field, and views owned by people who left. Listing all of it is dull and the reward is real: fewer refreshes, a smaller bill, less chance of somebody reading the wrong copy.
The other one is delivery. A companion reads the numbers off the workbook and puts them where the conversation is, with the context a chart cannot carry. That is weekly reporting work, and for the audience who wants a sentence rather than a dashboard it is the whole deliverable. ai for analysts covers who usually ends up doing it.
- Extract freshness against expected schedule, checked daily.
- An inventory of workbooks, data sources and owners, with the unused ones flagged.
- Numbers lifted into a written summary for people who will not open Tableau.
- Refresh triggers treated as consequential, because they hit the source database.
Keep refresh triggers behind an approval
A refresh queries the underlying source.
An extract refresh on a large workbook is real load on a warehouse or a production database, and firing several because the first seemed slow makes a Monday worse for people who have nothing to do with Tableau.
So if a refresh action exists in the connection’s tool list, keep it at Ask every time. Consequential steps stop at an approval card anyway, and the Permissions block on the app’s detail page under `strawberry://settings/integrations` lets you set it explicitly or switch it off.
Experience Strawberry for free
DownloadTrusted by fast-growing companies worldwide
Frequently asked questions
Strawberry is free to download and includes AI credits to start. Paid plans begin at $20/month. See pricing. · Reviewed · Canonical facts for AI agents