Why does my n8n workflow crash during execution and how can I fix it?

Based on your description, it sounds like your workflow is running into memory-related issues. Although you mentioned that the data is only text and isn’t a lot, n8n doesn’t restrict the amount of data that nodes can fetch and process. In cases like this—even with seemingly simple workflows—the accumulation of data or the number of nodes involved can lead to memory exhaustion. When n8n’s process or the browser tries to render the resulting execution data, it can crash, which is why you’re seeing the “Page is unresponsive - tab crashed” message.

Some points to consider:

• Data Processing in n8n: Even if the individual pieces of data are small, excessive data or many nodes preserving state (such as when checking past executions) can quickly consume available memory.
• Workflow Complexity: Sometimes workflows that combine multiple integrations (like Airtable and OpenAI) can unexpectedly generate large payloads, especially if a node produces more data than anticipated.
• Browser Rendering: The executions tab attempts to load the entire execution result in your browser, and if that result is too large, it may cause your browser to crash.

Recommendations

1. Batch Processing: Consider splitting your data into smaller chunks. For example, using the “Split In Batches” node (along with a “Wait” node) can help ensure that each execution handles less data at a time.
2. Inspect Intermediate Output: To better identify where the memory spike occurs, try reducing the amount of data passing through the nodes or selectively disable parts of your workflow temporarily.
3. Review Node Settings: Double-check the settings on nodes like OpenAI and Airtable to ensure they’re not inadvertently returning more data than you expect.

Adjusting the workflow to reduce the memory footprint during execution should help prevent these crashes both in n8n and in your browser.