Open Source Infrastructure Projects to Watch in

There's a particular kind of GitHub discovery that home lab people know well. You're scrolling Reddit at 11pm, you see a project mentioned offhand in a comment, you click it, you read the README, and somewhere around the third paragraph you think: why didn't this exist two years ago? That feeling—part excitement, part mild irritation at the gap that existed—is exactly the signal Brandon Lee of VirtualizationHowto says he now uses to filter what actually earns space in his infrastructure.

In a recent video, Lee walks through seven open source infrastructure projects he's been watching closely in 2026. The list is interesting not just for what's on it, but for the editorial framework behind it. Lee is explicit about his selection criteria: "I don't really want to deploy software just because it's new anymore. I want projects that actually solve problems that I have in my own home lab." That's a different bar than "interesting GitHub repo," and it shapes what makes the cut.

What emerges from the list is a rough portrait of where self-hosted infrastructure pain actually lives right now—and, implicitly, where the community has decided to show up and fix it.

The DNS problem nobody wanted to solve (officially)

Pi-hole HA leads the list, and the backstory is worth understanding. Pi-hole has been a home lab staple for years—network-wide ad blocking, local DNS, DHCP management. But high availability has always been the awkward footnote. The community built workarounds: Gravity Sync handled config synchronization for Pi-hole v5, and Nebula Sync picked up that role for v6. Both work, but they're bolted on externally.

Pi-hole HA attempts something more structurally ambitious: bringing configuration synchronization, DHCP failover, clustering, and virtual IP management into Pi-hole's own web admin console. Lee calls it "exactly what the Pi-hole community has been waiting for," and expresses hope that some of its ideas eventually make it upstream into Pi-hole proper.

That last part is worth sitting with. The history of community-built HA solutions eventually becoming official features is real, but it's not guaranteed. Plenty of excellent wrapper projects get stranded when the upstream project goes a different direction, or when the volunteer maintainers burn out. Pi-hole HA is genuinely filling a gap—what's less clear is whether that gap ever gets filled at the official level, or whether the community will be maintaining the bridge indefinitely.

The Docker management space keeps getting more interesting

Sencho is Lee's Docker Compose management pick, and he frames it against a crowded field—Portainer, Komodo, Arcane—that he describes as "great" before arguing that Sencho "approaches things a little bit differently." The differentiator he highlights is that Sencho understands your Docker Compose projects rather than just launching containers through a UI. Snapshotting compose configurations, an integrated app marketplace—it's positioning itself as a management layer that respects the compose-as-code workflow rather than abstracting away from it.

The Docker management space is genuinely competitive and arguably oversupplied. Portainer has enterprise backing and years of adoption. Komodo has its own enthusiastic community. Whether Sencho carves out durable space depends partly on how it handles the open source sustainability question—something the video doesn't address, and worth asking before committing your infrastructure to it.

Portbase is a different kind of bet. Originally a database backup tool, it's added Docker volume backup support—and that's what caught Lee's attention. The pitch is clean: containers are disposable, but the data in your volumes isn't, and Portbase uses native Docker volume access through local agents to back that data up to destinations including RustFS, Google Drive, or Azure Blob Storage. "Protecting those persistent volumes is going to be the increasingly important goal," Lee notes, and he's not wrong. As more people run meaningful workloads in containers, the backup story needs to grow up with them.

The governance backstory behind Incus

Incus is the project on this list with the most political backstory, and it's worth naming that clearly. When Canonical changed the direction of LXD—pulling it under its own commercial umbrella and shifting the open governance model—the Linux container community that had built around LXD as a neutral project had a problem. The community's response was Incus: a fork maintained under the Linux Containers project, developed by many of the same people who had built LXD.

Lee describes Incus as filling the "middle ground" between full VMs and Docker containers—the workload space where you need a complete Linux userspace but a full virtual machine is overkill. That's a real technical niche. But what makes Incus notable beyond the technical story is what it represents: a community successfully executing a fork when commercial interests changed the game. The VMware acquisition fallout Lee mentions as context has put a lot of organizations back in the market for trustworthy virtualization options, and "community-driven" has gone from feel-good marketing language to an actual evaluation criterion for a lot of infrastructure teams.

Kubernetes management, for humans

Omni from Sidero Labs addresses the Talos Linux Kubernetes management experience. Lee describes it as a centralized dashboard for cluster operations—upgrade Talos, update Kubernetes versions, manage lifecycle events from a single interface. His assessment of what it changed is striking: "It used to be a nightmare to even consider having to upgrade your Kubernetes cluster... Talos Linux has made that along with Omni as that management plane has made that super easy. It's almost a push button affair."

Kubernetes has a well-documented complexity problem, and the ecosystem response has generally been to add more tooling. Omni is interesting because it's tightly coupled to the Talos Linux opinionated approach—immutable, API-driven, no SSH—which means it inherits both Talos's operational advantages and its ecosystem constraints. If you're already in the Talos world, Omni seems to genuinely smooth the rough edges. If you're not, the entry cost is higher than just installing a management tool.

The networking and access layer

NetBird and Pangolin both address the question of how you connect to and expose your self-hosted services securely—from different angles.

NetBird combines WireGuard with identity-based zero trust access control. Rather than treating networking as a layer that connects everything to everything, it grants access based on identity and policy. Lee's argument here is structural: "With a traditional VPN, you're often exposing much more of your network than you really need to." Zero trust principles aren't new, but open source implementations that don't require enterprise contracts are still relatively young. NetBird's integration with identity providers positions it as something that can grow from home lab experiment to serious infrastructure.

Pangolin is the most explicitly "in progress" project on the list. The problem it targets—securely exposing self-hosted services to the internet without stitching together reverse proxies, SSL cert management, DNS, firewall rules, and auth across multiple dashboards—is genuinely painful. Lee is honest that Cloudflare Tunnel, Tailscale, and others already exist in this space, and he frames Pangolin as "experimenting with a different workflow" rather than claiming it's already won. If it matures in the direction Lee expects, it could make secure service publishing dramatically less time-consuming.

What the list adds up to

Taken together, these seven projects sketch a particular moment in self-hosted infrastructure. The tooling is maturing past "deploy things" toward "operate things reliably"—more HA, better backup, less manual lifecycle management, cleaner access control. That's a sign of an ecosystem growing up, which is mostly good news.

But the home lab community operates in constant negotiation with sustainability realities that don't apply to enterprise software. Several of these projects are early. Some depend on small maintainer teams. Pi-hole HA hopes its ideas become official features; Incus proved a fork can succeed but also showed how much energy that takes; Sencho and Pangolin are still establishing themselves.

Lee's framing—"our time is one of the most valuable commodities that we have"—is a useful lens for evaluating any of these. The question isn't just whether a project solves a real problem today. It's whether it'll still be solving it in two years, with an active community, when you need to upgrade your infrastructure at 11pm because something broke.

That answer isn't in any README. It rarely is.

Dev Kapoor covers open source software and developer communities for Buzzrag.