Add schedule capability to the Evaluation Trigger

Feature Requests
1

The idea is:

Currently the Evaluation Trigger can only be fired manually. In order to have a periodical and automated checking of AI Agent drift (e.g., daily or weekly) it would be nice to make the Evaluation Trigger run on scheduled basis.

My use case:

There is a requirement to have a daily Canary Test against several AI Agent workflows to ensure that no context rot or quality drift that degrades AI Agent perf silently. The Canary Test is something similar to this: Monitor AI quality drift with GPT-4o-mini evaluations and Slack alerts | n8n workflow template.

I think it would be beneficial to add this because:

While I can click the Evaluation Trigger manually or maybe use the Playwright to automate the clicking, making the Evaluation Trigger schedule-able can make things much easier and cleaner (no human clicks, no Playwright setup).

Any resources to support this?

Are you willing to work on this?

Yes.

I have developed a set of custom nodes and have understanding about how an generic n8n Trigger-type node works. I will take some time to study the internals of Schedule Trigger and see how to enrich the Evaluation Trigger node.

2

Hi everyone, :waving_hand:

Thanks @Victor_Tseng for the great insights on this thread! :folded_hands:

First off, I wanted to vote for this feature request, but I’ve unfortunately reached my vote limit for the moment! @Victor_Tseng make sure you’ve voted for your own topic at the top of the page to help it gain more visibility.

I’ve been thinking about how to bridge the gap for scheduling these evaluation triggers. Initially, I considered developing a Chrome extension to handle this. It seems like a quick solution on the surface, but after looking at the long term stability needed for n8n workflows, I’ve decided I don’t recommend that route.

Instead, I want to suggest using Selenium (Python). While an extension is a solution, I believe Selenium is much better for a professional setup. Here is why I’m focusing on it:

1. Stability and Independence

A Chrome extension usually requires an active browser window and a GUI to be running. If the worker’s computer restarts or the browser crashes, the trigger dies. Selenium can run in headless mode on a server or within a Docker container. It isn’t affected by “active worker” status—it just runs reliably in the background.

2. Deep n8n Integration - CLI

Selenium scripts can be integrated directly with the n8n CLI or called via the Execute Command node. This allows you to trigger evaluations as part of a larger server side process rather than relying on a client-side browser plugin.

3. Centralized Management

Instead of managing extension settings across different browsers, you can build a Centralized Evaluation Hub. One Selenium script can be designed to test and trigger multiple different workflows, making it much easier to scale as your automation grows.

4. Advanced Monitoring and DB Storage

This is a huge advantage. When using Selenium with Python, you can:

  • Store all evaluation results and logs in a Database (Postgres - MySQL).
  • Use n8n to monitor those DB entries to trigger alerts.
  • Build custom retry logic and error handling that a simple extension just can’t handle.

5. Security and Governance

From a security perspective, running your own code on your own infrastructure is always safer. You don’t have to worry about the permissions or data sharing risks often associated with browser extensions.

I know there are modern alternatives like Playwright or Puppeteer that follow this same logic, but my experience is primarily with Selenium, and it has proven to be incredibly robust for this “business layer” of automation.

I’d love to hear if anyone else has moved away from browser based hacks in favor of server side scripts for their triggers!

2 Likes