Currently, I am using self-hosted harbor repo for the docker repository.
Due to some reasons, I must migrate away from the above harbor repo.
Currently, most/all of our infra is in GCP, so I was thinking of migrating to GCR.
There are a few blockers with the GCR:
- it doesn't provide information, like when was the last time a certain image was pulled. I need this information, to clean up images which haven't been used in the last 1 month. Currently, I get all this data from harbor repo, and using a script, can clean older images.
Some solutions:
- Manage the list of images being pulled in GKE, and use this for cleanup.
- Use a middleware or proxy to GCR, which can provide me with this data. For this, I looked into harbor. And this harbor can run on the same GKE cluster and will work as a proxy cache too. But Harbor will keep a copy of each of the images, it proxies, and even older images as well, i.e. it will work as a persistent repo, and not just a cache. I don't want to do that.
Is there any better way to do this? Or any other software/tool, which I can use to implement this. Any suggestions would be helpful. Thanks!