Performance improvement with Buffered Logging? 🤔

[pbdco]

Hi n8n community! We’ve submitted a feature request for a buffered logging system that could significantly improve workflow throughput. Currently, direct database writes for execution logs create a bottleneck (~30 req/s limit), from what we are seeing in some performance benchmarks in clustered setups in k8s.*

What’s the proposal?

• Implement buffered logging to batch database writes
• Configurable at workflow and instance level
• Optional Redis support for distributed setups
• No compromise between performance and logging capabilities

Why is this important?

• Improves workflow throughput
• Maintains comprehensive logging for compliance/debugging
• Works with existing queue mode architecture

If you’re running high-throughput workflows and want better performance without sacrificing logging, please upvote and share your thoughts!

Vote this feature here:

Buffered Logging for Performance Improvement

The problem is: Direct database writes for execution logs in n8n create a significant performance bottleneck, limiting workflow throughput despite the number of workers and webhook instances. Users must choose between having logs (with limited performance) or disabling logs entirely (losing crucial operational visibility). The solution is: Implement optional buffered logging to batch database writes, similar to how Nginx and other high-performance systems handle logs. This allows maintaining c…

*Related discussion: N8N benchmarking throughput - max 30 req/s?

Let’s make n8n even better together

[n8n]

It looks like your topic is missing some important information. Could you provide the following if applicable.

• n8n version:
• Database (default: SQLite):
• n8n EXECUTIONS_PROCESS setting (default: own, main):
• Running n8n via (Docker, npm, n8n cloud, desktop app):
• Operating system:

[gualter]

hi @pbdco

Thanks for raising this. Our product team carefully reviews any feature requests we see from the community so thank you for your collaboration