Hi everyone,
I’ve encountered an unexpected behavior when using the BigQuery node with the “Dry Run” option enabled, and I wanted to check if this is the expected behavior or a potential issue.
Here’s what’s happening:
- I enable the “Dry Run” option in the BigQuery node.
- When the workflow executes, the node returns the number of bytes to be processed, as expected.
- However, it also returns the actual results of the query, which I wasn’t expecting in a dry run.
- Additionally, when I check the BigQuery UI, the query is marked as billed, which defeats the purpose of a dry run.
For comparison, if I use the HTTP node and call the BigQuery API manually with dryRun: true, the behavior is different:
- The query is not billed.
- The query results are not returned, which aligns with the expected dry run behavior.
So I’m wondering:
- Is this the intended behavior for the BigQuery node in n8n?
- Or could this be an issue or a misconfiguration on my end?
This could have implications, especially now that BigQuery is available as a tool for AI Agents, since you may not want AI Assistants triggering expensive queries unintentionally. I’m aware of the “maximum bytes billed” setting as a safeguard, but it still feels like dry run should behave differently.
Thanks in advance for your help!
Here is a query you can try in “dry run” with a public dataset. It should be billed approx 1.43 MB, even with the dry run enabled, so be aware of this if you try it 
Information on your n8n setup
- n8n version: 1.89.2**
- Database: SQLite
- n8n EXECUTIONS_PROCESS setting: own, main
- Running n8n via: Docker