Intermittent BigQuery job failures using Workload Identity Federation in EMEA
Sep 24, 13:27 UTCSep 25, 01:50 UTC
Duration
12h 23m
Impact
Minor
Root cause
Not disclosed
dbt Labs, 90 days
28 incidents
Affected
Not listed by the vendor.
Final update
The issue has been resolved, and all affected systems are now functioning normally. Please contact Support via email [support@getdbt.com](http://support@getdbt.com "support@getdbt.com") if you continue to experience delays and are unsure of the root cause.
Timeline
- Resolved · Sep 25, 01:50 UTC
The issue has been resolved, and all affected systems are now functioning normally. Please contact Support via email [support@getdbt.com](http://support@getdbt.com "support@getdbt.com") if you continue to experience delays and are unsure of the root cause.
- Monitoring · Sep 25, 00:43 UTC
The impacted component appears to be recovering, we will continue to monitor the performance. Please contact us at support@getdbt.com with any questions or concerns.
- Identified · Sep 24, 20:59 UTC
We have identified an issue with intermittent failures for jobs that connect to BigQuery using Workload Identity Federation. If you're affected, please open a case with Google Cloud support about the 503 errors from the Workload Identity Federation token exchange. Once you receive more information from Google, you can share it with us by opening a dbt Support ticket so we can investigate further. We will provide further updates as soon as new information becomes available.
- Monitoring · Sep 24, 20:21 UTC
We have not observed any further occurrences and are monitoring to confirm the issue has fully cleared. dbt job runs have returned to their normal state and we are continuing to monitor.
- Investigating · Sep 24, 18:41 UTC
We're continuing to investigate intermittent failures for jobs in the EMEA multi-tenant region (AWS) that connect to BigQuery using Workload Identity Federation. If you're affected, please open a case with Google Cloud support about the 503 errors from the Workload Identity Federation token exchange. Once you receive more information from Google, you can share it with us by opening a dbt Support ticket so we can investigate further. We will provide further updates as soon as new information becomes available.
- Investigating · Sep 24, 17:32 UTC
We're continuing to investigate intermittent failures for jobs in the EMEA multi-tenant region (AWS) that connect to BigQuery using Workload Identity Federation. If you're affected, please open a case with Google Cloud support about the 503 errors from the Workload Identity Federation token exchange. Once you receive more information from Google, you can share it with us by opening a dbt Support ticket so we can investigate further. We will provide further updates as soon as new information becomes available.
- Investigating · Sep 24, 15:19 UTC
We're continuing to investigate intermittent failures for jobs in the EMEA multi-tenant region (AWS) that connect to BigQuery using Workload Identity Federation. Affected runs fail with "The service is currently unavailable" (HTTP 503) during the token exchange with Google Cloud. We haven't found any changes on our side, and the error appears to originate on Google Cloud's side. If you're affected, please open a case with Google Cloud support about the 503 errors from the Workload Identity Federation token exchange. Once you receive more information from Google, you can share it with us by opening a dbt Support ticket so we can investigate further. We will provide further updates as soon as new information becomes available.
- Investigating · Sep 24, 13:27 UTC
We're investigating an issue with BigQuery authentication that is causing some job runs to fail intermittently. This is impacting jobs in the EMEA multi-tenant region (AWS) that connect to BigQuery using Workload Identity Federation. Affected runs fail with "The service is currently unavailable" (HTTP 503) while authenticating with Google Cloud, before any queries run. The team is working on a resolution and we will provide updates as soon as new information becomes available. In the meantime, rerunning a failed job or using Rerun from failure will usually succeed. Jobs using other BigQuery authentication methods, jobs on other data platforms, and dbt Studio are not impacted and should be running as expected.
More from dbt Labs
Full history| Started | Incident | Impact | Duration |
|---|---|---|---|
| Sep 2423:56 UTC | GitLab service disruption | minor | Ongoing |
| Sep 2322:59 UTC | Intermittent dbt deps failures | minor | 4m |
| Sep 1823:31 UTC | Schema Hydration Panic | major | 14h 58m |
| Sep 1014:28 UTC | Intermittent errors on the Discovery API | minor | 2h 17m |
| Sep 1013:46 UTC | Increased error rates associated with dbt MCP and dbt Wizard | minor | 41m |
| Sep 1008:02 UTC | State usage metrics not being ingested | major | 1h 13m |
Sources: vendors' own status pages, published postmortems and SEC 8-K Item 1.05 filings, read daily. Times as reported. Logos via logo.dev; trademarks belong to their owners.