We lost 3800 stars on Github in 1 click đ
Never switch public repository to private
Our team leaked an API key in a Git commit during maintenance and was immediately alerted by GitGuardian. To revert that change and clean up the git history, Pierre decided to switch our repository from public to private until our team fixed it. Never do that! We just lost all the stars and the forks of our repo when switching back to public đ.
The Github support
Well, we desperately tried to contact GitHub support. And we receive this response...
Please, help us!
We face an unusual situation, and we ask for your help! There are two ways of helping:
- I am sure Github can do something about that, but they don't want to bother. Can you share this story with your friends working at Github? If someone working there can help, they can contact us via hello {at} qovery {dot} com. I will send you a super Qovery swag pack with a great bottle of French red wine đ· (I promise it will be excellent!).
- If you like what we are doing at Qovery - please give us a star âïž. You can read this Reddit post when I announced the open-source launch of our Qovery Engine.
I hope this post will help you never make the same error as we did!
---
Update on the 6th of December 2021
The community is lovely đ€© Thank you all for your fantastic support. We received support from people all around the world. It is crazy. Thank you, thank you, thank you. We almost reached 1k stars again - please keep supporting us with a star âïž. GitHub is also considering this incident, and I feel they tried their best to solve our issue.
.svg)
.svg)
.svg)
Suggested articles
The cluster coming up is the easy part. What catches teams off guard is what happens six months later: certificates expire without a single alert, node pools run at 40% over-provisioned because nobody revisited the initial resource requests, and a manual kubectl patch applied during a 2am incident is now permanent state. Agentic control planes enforce declared state continuously. Monitoring tools just report the problem.
The instinct when setting up Kubernetes observability is to instrument everything and send it all to your APM vendor. That works fine at ten nodes. At a hundred, the bill becomes a board-level conversation. The less obvious problem is the fix most teams reach for: aggressive sampling. That is how intermittent failures affecting 1% of requests disappear from your monitoring entirely.
Scaling your deployments to zero is only half the battle. If your cluster autoscaler does not aggressively bin-pack and terminate the underlying worker nodes, you are still paying for idle metal. True environment sleeping requires tight integration between your ingress layer and your node provisioner to actually realize FinOps savings.
The structure, table, tool list, and code blocks are all worth keeping. The main work is fixing AI-isms in the prose, updating the case study to real metrics, correcting the FAQ format, and replacing the CTAs with the proper HTML blocks. The tool descriptions need the "Core strengths / Potential weaknesses" headers made less template-y, and the intro needs a sharper human voice.
For years, Red Hat OpenShift has been the safe choice for heavily regulated, on-premise environments. It operates as a secure fortress. But in the public cloud, that fortress acts as an expensive prison. Paying proprietary per-core licensing fees on top of your standard AWS or GCP compute bill is a redundant "middleware tax." Escaping OpenShift requires decoupling your infrastructure from your developer experience by running standard, vanilla Kubernetes paired with an agentic control plane.
Agentic Kubernetes resource reclamation is the practice of using an autonomous control plane to continuously identify, suspend, and delete idle infrastructure across a multi-cloud Kubernetes fleet. It replaces manual cleanup and reactive autoscaling with intent-based policies that act on business state, eliminating the configuration drift and cloud waste typical of unmanaged fleets.
Kubernetes focuses on container orchestration, but the reality on the ground is far less forgiving. Provisioning a single cluster is a trivial Day-1 exercise. The true operational nightmare begins on Day 2. Teams that treat multi-cloud fleets like isolated pets inevitably face crushing YAML configuration drift, runaway AWS bills, and severe scaling bottlenecks.
.webp)