Proxmox Notifications: Setting Up Alerts for Backups, Replication, and Node Health Before Something Breaks Silently

Every other article on this site that mentions a scheduled Proxmox job, backups, replication, HA, eventually says some version of “check the GUI periodically to make sure it’s still running.” That’s fine advice until the day you stop checking, which is also usually the day it matters most. Proxmox has had a real, built-in notification system since 8.1 that can push a failed backup, a broken replication job, or a flapping node straight to email, a webhook, Slack, Discord, or whatever else you already watch, and most homelabs never turn it on past whatever came configured by default. This is the setup that closes that gap. ...

November 2, 2026 · 7 min

Homepage vs Homarr vs Heimdall: Picking a Self-Hosted Dashboard for Your Homelab

At some point every homelab operator ends up with a browser bookmarks folder that’s really a map of their own infrastructure: Sonarr on one port, Proxmox on another, Uptime Kuma somewhere else, each one a separate tab to dig up when you actually need it. That works fine for three or four services. It stops working once you’re running a dozen, especially if anyone else in the house wants to check on something without memorizing 10.0.x.x:port combinations. A self-hosted dashboard solves exactly that problem: one page, usually set as your browser’s homepage or a wall-mounted tablet’s only tab, that links out to everything and optionally shows you live status while it’s at it. Homepage, Homarr, and Heimdall are the three worth seriously considering, and they take meaningfully different approaches to the same job. ...

November 1, 2026 · 7 min

Choosing a Server Chassis: 1U vs 2U vs 4U, Airflow, and Drive Bay Layout

The rack article on this site covers the cabinet: depth, rails, PDU, cable management. This one covers the thing that actually goes into the rack, the chassis itself, and it’s a separate decision with its own traps. A rack sized right for a server that turns out to have the wrong airflow direction, too few drive bays, or a riser that won’t clear a double-width GPU is still a wasted purchase, just one layer in from the one the rack article warns about. ...

October 30, 2026 · 7 min

Proxmox Storage Replication: Near-HA Without Ceph or Shared Storage

The clustering article on this site already made the point that shared or replicated storage is the real gate on HA failover, not quorum. What it didn’t do is explain the replication half of that sentence, because it’s a separate feature with its own setup, its own limits, and its own failure modes. If you’re running a two or three node homelab cluster without Ceph and without a NAS acting as shared storage over NFS or iSCSI, Proxmox’s built-in storage replication is very likely the tool that gets you automatic failover anyway, and it’s worth understanding on its own rather than as a footnote to HA. ...

October 29, 2026 · 8 min

Self-Hosted Team Chat: Matrix vs Rocket.Chat vs Mattermost for Your Homelab

Every few years another wave of people gets tired of a chat platform tightening its terms of service, shutting down a client, or just getting bought by someone they don’t trust, and goes looking for a self-hosted replacement. The three names that always come up are Matrix, Rocket.Chat, and Mattermost. They get lumped together as “self-hosted Slack alternatives,” but they’re not really competing for the same job. Picking wrong means either running something that fights you every day or running something with ten times the operational weight you actually need. ...

October 28, 2026 · 7 min

PCIe Bifurcation and M.2 Expansion Cards: Getting More NVMe Slots Out of a Consumer Motherboard

The CPU and motherboard article on this site already covers how PCIe lane budgets get tight on a consumer platform once a GPU, an HBA, and a NIC are all fighting for slots. This piece is about the thing people reach for once that budget is spent: PCIe expansion cards that turn a single slot into multiple M.2 NVMe slots. Some of these cards work in any PCIe slot on any board. Others need a specific motherboard feature called bifurcation, and buying the wrong card for your board is a common, entirely avoidable way to end up with an expensive piece of plastic that Windows or Linux only sees one drive on. ...

October 26, 2026 · 7 min

Proxmox SDN: When VXLAN and EVPN Zones Actually Beat a VLAN-Aware Bridge

Proxmox has had a Software Defined Networking (SDN) subsystem built into the web UI for several versions now, living under Datacenter > SDN, and most homelabbers never open that tab. That’s reasonable: a VLAN-aware bridge and a managed switch trunk port solve network segmentation for a single-node or small-cluster homelab just fine, and that’s the setup covered in most Proxmox networking guides, including the basic bridges-and-VLANs approach. SDN solves a different, narrower problem, and it’s worth understanding what that problem is before reaching for it, because it adds real operational complexity that isn’t justified unless you actually have the problem it’s built for. ...

October 25, 2026 · 7 min

k3s vs k0s vs MicroK8s: Do You Actually Need Kubernetes in Your Homelab?

Kubernetes shows up in homelab conversations constantly, usually right after someone says “I want to get better at this for work.” That’s a legitimate reason to run it. What’s less legitimate is the assumption that Kubernetes is a natural next step up from Docker Compose the way Compose was a natural next step up from a pile of docker run commands. It isn’t. Compose and Kubernetes solve overlapping problems with very different amounts of machinery, and for a huge share of homelab workloads, Compose (or a UI on top of it, like Portainer or Dockge) is simply the correct tool, not a beginner’s tool you’re supposed to graduate out of. This is about when a lightweight Kubernetes distribution actually pays for itself at home, and when it’s complexity you’re taking on for its own sake. ...

October 24, 2026 · 8 min

Choosing a Network Interface Card for Your Homelab: Chipsets, Drivers, and Why Not All NICs Are Equal

The 10GbE and SFP+ articles on this site both assume you already have a network card worth plugging something into. That’s not always true. Plenty of homelab trouble that looks like a cabling problem, a switch problem, or a flaky driver is actually a bad NIC choice from the start, usually an onboard chipset that was never meant to run under constant load with a hypervisor’s network stack sitting on top of it. This is the piece that covers the card itself: which chipsets to trust, which to avoid, and what actually matters when you’re picking an add-in NIC for Proxmox, TrueNAS, or any other self-hosted box. ...

October 22, 2026 · 7 min

Hardening Proxmox: Locking Down SSH, the Web UI, and Root Access on Your Homelab Host

Every other security decision in your homelab is scoped to one service. If Vaultwarden gets popped, you lose your passwords. If Jellyfin gets popped, someone can watch your movies. If your Proxmox host gets popped, the attacker owns every VM and LXC running on it, plus whatever network access those guests have. The hypervisor is the one layer where “harden it properly” isn’t optional homelab polish, it’s the thing that determines whether a bad day stays contained to one service or takes out the whole stack. ...

October 21, 2026 · 7 min