Lynq
Turn database rows into running Kubernetes apps
Start new thread

Lynq - Turn database rows into running Kubernetes apps

by•
Lynq is an open-source Kubernetes operator that turns your database into the source of truth for infrastructure. Every active row becomes a fully provisioned, continuously reconciled stack. add a row to deploy, deactivate it to clean up. Apache 2.0.

Add a comment

Replies

Best
Maker
📌
Hi Product Hunt, I'm Tim. I build Lynq, an open source Kubernetes operator. The idea is simple. If you run a multi-tenant SaaS, your database already knows exactly which customers exist and which ones are active. But the infrastructure for those customers is provisioned somewhere else, by scripts, pipelines, or a person. Those two views drift apart. The database says 200 active tenants, the cluster has 198 and a half, and nobody notices until something breaks. Lynq removes that gap. It watches your database and reconciles every active row into real Kubernetes resources. Insert a row and a full stack comes up. Set activate to false and everything is torn down cleanly. There is no pipeline to trigger and no script to babysit. Some details I care about: - Reconciliation is continuous. Drift gets corrected all the time, not just at deploy time. - 16 resource types are supported natively. Raw manifests cover everything else. - Lifecycle policies control creation, deletion, and conflicts. You can retain customer data on offboarding instead of deleting it. - Prometheus metrics, alert rules, and a Grafana dashboard ship with it. - The quickstart takes about five minutes. One script sets up a local cluster, MySQL, and sample resources. Lynq is Apache 2.0 licensed and a solo project. Issues and feedback have a direct effect on what I build next. I would love to hear how you provision per-customer infrastructure today, and what datasource you would want supported next. Docs: GitHub: Thanks for reading.

The database-as-source-of-truth idea is so obvious in hindsight and im surprised more teams arent doing this. Reconcile loop against SQL rows feels way cleaner than juggling a pile of CRDs.

Maker

 Thank you so much! I really appreciate you taking the time to share this.

That was exactly the motivation behind Lynq. As a solo developer, I wanted to keep the simplicity and reliability of the Kubernetes reconciliation pattern without introducing lots of CRDs for application state. Using a SQL database as the source of truth felt like a much more natural fit.

Your comment is really encouraging. Thanks again!

Finally, a Kubernetes operator that doesn't make me babysit deployments. Loved watching a row flip and seeing the full stack reconcile itself without a single manual apply.

Maker

 Haha, thank you! I'm really happy to hear that. 😄

That was exactly the experience I was hoping to create/change a row, let the reconciler do the rest.

Feel free to experiment with it however you like! If you find an interesting use case, run into any issues, or have ideas for making Lynq better, I'd love to hear about them.

Database-as-config approach is super clean, made a test deployment in minutes just by inserting a row. Reconciliation loop feels solid and the open source license is a nice touch.

Maker

 Thank you! I'm really glad to hear you had a good experience.

I've actually been running Lynq in production for over two years, where it manages more than 300 LynqNode resources every day. It has proven to be a reliable approach for my own use case, but I'm even more excited to see how others apply it to different problems.

If you end up using Lynq in your own projects, I'd love to hear about your use case, any tips you've discovered, or areas where you think it could be improved. Community feedback is one of the best ways to make the project better.