Having this same issue. No changes made but having issues across various apps and implementation types now. After going through all of my logs I was able to pinpoint my first failed job on one query to
First failure: 2026-03-12T03:12:23 UTC
Last success: 2026-03-11T02:17:43 UTC
Sometime between those two timestamp some update seems to have been made causing this failure. Looking through my logs further I noticed on the bigquery side there is a change in how my “named parameters” are showing. For my specific implementation I used to get @=23791571792 which worked well, now I see @={"RT$BQ_TYPE":"INT64","value":23791571792} which is causing errors.
I seem to have fixed this by changing what used to exist on the retool query field
But as the initial post says, I have not changed anything on my apps related to these update queries in a long time. I don’t remember why I had this one specific field setup to insert with {RT$BQ_TYPE: \'INT64\',.... but not on my other fields. So the real question might be, was there a bug that got patched now causing my workaround to fail, OR is this a new bug.
We recently made a change that has since been disabled for all Cloud users that should have fixed this issue. Can you confirm that your BigQuery resource is now working again?