You might say: “Wasn’t this possible with Kind?”. In theory - yes. In practice that meant DinD and a lot of friction. vCluster takes a smarter path.

There is already a huge number of alternatives to Kubernetes tooling for every purpose - especially when it comes to observability 📈 - from full-stack platforms to simple add-ons, open source or enterprise. I know I won’t shock anyone with statements like “live event analysis” or “eBPF-based service maps”. I’ll even abstain from words like “better”, “more” or “faster” - I’d rather say it’s simply a good alternative.

Anyone who has ever built a DIY database setup on Kubernetes knows the pain 🤦‍♂️. Yes, operators from Zalando, CNPG, StackGres, Percona and others have already made life much easier - but you still end up stitching different APIs, UIs, CRDs and workflows together depending on which engine you run.

Dragonfly (d7y) officially became a CNCF Graduated project on January 14, which makes this a perfect moment to talk about it. Dragonfly is an open-source, P2P-based file distribution and image acceleration system - essentially a private BitTorrent for cloud-native environments.

I’ve been running a self-hosted Vaultwarden for almost 4 years for personal use and I’m quite happy with it: it supports granular access management, item scoping, has a web UI, is fully compatible with the official Bitwarden mobile apps (Android & iOS) and even the Terraform provider.