While executing Google Sheets API node (Resource: Sheet Within Document, Operation: Append Row, Mapping Column Mode: Map Each Column Manually) n8n takes too much memory (from 20% to 50% of 1 GB RAM on one manual execution). Sometimes node executes too long (about 20 minutes or so). But it’s just adds data without long strings. What could be the problem? Thanks in advance for answer.
1- Fragment the work (“batch processing”)
Use Loop Over Items or SplitInBatches nodes to process rows in small chunks (e.g., 200 rows per batch) instead of loading the entire sheet. Each batch is reduced, processed, and freed up memory before the next.
2- Separate into sub-workflows. Create sub-workflows executed by Execute Workflow; each one works with a batch and returns reduced results. This keeps the memory footprint per execution low.
3- Avoid manual execution for heavy tasks. Use cron or webhook triggers instead of the “Run” button in the UI to avoid maintaining large data sets on the frontend.
4- Increase available memory. If you are limited by the container or server’s RAM, allocate more (2–4 GB), or on n8n Cloud, consider a plan with higher RAM.
5- Optimize the node mode. If you only add rows, consider using Google Sheets V1 or an interface that doesn’t require loading the entire sheet. In V2, limit the read range if you don’t need to download all the data.
Thank you for your reply. Yes, Google Sheets API v2 does indeed use a lot of RAM. In addition, I couldn’t find an option to limit the sheet load. In my workflow, I replaced the built-in nodes with HTTP Request with Google Sheets API. It no longer consumes so much RAM.