TE
TechEcho
Home24h TopNewestBestAskShowJobs
GitHubTwitter
Home

TechEcho

A tech news platform built with Next.js, providing global tech news and discussions.

GitHubTwitter

Home

HomeNewestBestAskShowJobs

Resources

HackerNews APIOriginal HackerNewsNext.js

© 2025 TechEcho. All rights reserved.

Show HN: I Built a GitHub Action to Monitor LlamaIndex Performance

2 pointsby Ephil012over 1 year ago
Hi all, I had been working with tools like LlamaIndex at work and realized that there were not many good options for monitoring RAG systems for performance. So I built a pretty simple Github Action + PyTest setup to measure the performance of LlamaIndex over time. The setup basically uses a library called Tonic Validate, which scores the quality of your RAG system’s answers (Disclaimer: My current company made Tonic Validate and I am an engineer on Tonic Validate). Using Tonic Validate, it scores the responses from LlamaIndex on a set of test data I created and then uploads it to Tonic Validate’s UI for visualization. If anyone is interested in it, you can find the full source code here [1]. I also wrote a guest blog post on LlamaIndex’s blog about it here [2] if anyone wants more details about how it works.<p>1. https:&#x2F;&#x2F;github.com&#x2F;TonicAI&#x2F;llama-validate-demo<p>2. https:&#x2F;&#x2F;blog.llamaindex.ai&#x2F;tonic-validate-x-llamaindex-implementing-integration-tests-for-llamaindex-43db50b76ed9

1 comment

gloria305over 1 year ago
Hmm, couldn&#x27;t quite get the use of it - I did see it&#x27;s a 100% python - maybe use pypi package for it instead?