Grai is the open source data observability tool that catches bugs before they make it to prod. Using data lineage it understands how changes impact your data stack and reports back version control. It even runs dbt tests against prod data migrations.
Hey PH! Edward and I started building Grai because existing open source data observability tooling was all fundamentally reactive. We wanted something developers could use to prevent outages before going to production rather than waiting and alerting us when something broke.
We use the tool internally and hope it’s useful to other engineers as well 😀.
Report
This sounds very useful. Any tips on setting up and configuring a new deployment? Some best practices you've learned from using it as you've been building it?
@bentley_cook Absolutely! Ease of setup is actually one of our biggest differentiators . You don't have to change anything about your stack and can get started in 2 steps.
1. Add data connections to your various sources through the web app. This usually just requires adding read credentials to your database or uploading a file like dbt
2. Add one of our pre-built Github Actions workflows wherever you deploy code. This is about three lines of additional code and you can find more info in the docs: https://docs.grai.io/tooling/git...
After that you're setup and ready to go, your action will run on pull requests evaluating data changes for impact on the rest of your infrastructure.
Report
Congratulations on the launch!! 🚀 love the team & execution, best of luck moving forward!
🔑 preventing outages >> fixing outages
Currently not, however this is a feature we will consider. We do have a GitHub action and a GitHub App https://github.com/apps/graibot which allows you to connect Grai to GitHub.
Grai
Grai
Grai
Grai
Vizzly
DiffSense
Grai