Why it matters
When a refresh fails, nothing visibly breaks. The dashboard still opens, but the numbers are out of date, and the first person to notice is often a business user who has already made a decision with them. Over time, repeated failures erode trust in the whole platform.
The usual causes
1. Credentials that changed or expired
A password rotation, an expired token or a person leaving the company can break every refresh that depended on their credentials. Embedded personal credentials are the most common culprit.
2. The source changed
A renamed column, a removed table, a moved file or a changed path in the source system will stop a refresh that worked yesterday.
3. Queries that take too long
Large extracts, slow sources and complex custom SQL can run past the time limits your administrators have set.
4. Too many jobs at once
When many refreshes are scheduled for the same busy window, jobs queue, run late and are more likely to fail. See why Tableau refreshes run late.
5. Connectivity
On Tableau Cloud, refreshes of on-premises data depend on Tableau Bridge being available. Network changes and firewalls can cause failures on Tableau Server too.
6. Suspended schedules
Tableau can suspend a schedule that keeps failing, or one for content nobody uses. A suspended refresh stops failing because it stops running. See why Tableau suspends extract refreshes.
Where to look
Tableau records the outcome of every refresh. On Tableau Server, the built-in administrative views show recent background tasks and their errors. On Tableau Cloud, site administrators can review jobs and refresh history for the site. The error message usually points to the cause.
Focus on patterns, not single failures
A single failure is often temporary: a source was briefly unavailable, or a network blip occurred. The failures that deserve attention are the ones that keep happening to the same content, or that start right after a change. Fixing those few usually removes most of the noise.
Make the fixes stick
- Use service accounts for important data sources rather than personal credentials.
- Agree change control with source system owners, so schema changes are planned, not discovered.
- Make extracts lighter: incremental refreshes, filters, aggregation and hiding unused fields all help.
- Spread schedules out of the busiest windows.
- Turn on failure notifications so owners hear about problems first. Tableau documents how in Notify owners when extract refreshes fail.
- Keep ownership current when people change roles or leave.
Common questions
Why do Tableau extract refreshes fail?
Most often because credentials changed or expired, the source changed, queries took too long, too many jobs ran at once, connectivity was lost, or the schedule was suspended.
How do I find failed refreshes?
Tableau Server’s built-in administrative views and Tableau Cloud’s site job history show recent refresh outcomes and error messages.
How do I stop the same refreshes failing again?
Fix the cause rather than re-running the job: use service accounts, plan source changes, lighten extracts, spread schedules and make sure owners receive failure notifications.
Further reading: Tableau Help, Notify owners when extract refreshes fail. Tableau and Salesforce are trademarks of Salesforce, Inc.