All Articles

Bitnami Docker Repository Changes Broke My Deployments

Last week started like any other. I pushed a small update to one of my services, hit deploy… and suddenly everything blew up. My Kubernetes jobs failed, Helm upgrades didn’t go through, and console was full of errors about missing Docker images.

After a bit of digging, I realized what had happened - Bitnami had changed their Docker repository policy and naming convention. Many images that used to live under the old bitnami namespace were moved into bitnamilegacy registry and anything else than latest became unavailable. That meant charts, but also new nodes without cache were trying to pull images that no longer existed.

Bitnami Docker Repository Change Broke My Deployments

For example, this classic line docker pull bitnami/postgresql:17 just stopped working.

It wasn’t just me - the issue spread quickly across GitHub issues, forums, and Reddit. Countless pipelines broke overnight because of this change. This change was “introduced” by Bitnami during summer without much warning, and many users were caught off guard. New model was forcing teams to use only latest tags or move to paid plans for other versions - which is not feasible for many open-source projects or small teams, and not good for production.

Hey, it’s still broken!

Luckily, the hotfix wasn’t too bad once I figured it out. First I switched my deployments to use official legacy images (to allow project to run smoothly for now), and started to look for new community-maintained alternatives.

The cool part? The community’s already rallying around new initiatives - new Docker repos, forked Helm charts, and guides for migrating away from Bitnami. It’s great to see how fast open-source communities can react when something like this happens. This is the new opportunity to be more diverse in tooling choices - and not be locked again into a single provider. Of course, a lot of this new charts are not perfect - but they grow fast with community help. It’s sometimes hard to replace something that was usually used as a cornerstone, and for a simple implementation “it just worked”.

For now, I saw the shift in the helm charts I use - all the solutions start using original images, and are more dependent on the init scripts than specially built images by third-party provider like Bitnami. I guess that vendor-locking and keeping code for yourself - specially for basic images and projects - destroyed a trust in this field for many people, and going to move in different way.

Even now if you look for images like redis on ArtifactHub, for now Bitnami is on the top - but many alternative charts are already gains popularity. Actively maintained, truly open-sourced, and with community support. For now, it looks like future is bright.

Whats now?

What I can say, for sure lesson was learned. Even if something has “just worked” for years, it’s worth keeping an eye on what dependencies you rely on. One namespace change can bring a lot of things down fast.

To the next update!