SysAdmin Weekly #23: Build Something Nobody Asked For
The queue will be there Monday. Your curiosity might not be.
TL;DR
Was at a conference recently. Half the room was on their phones. That’s not a people problem.
The culture of IT quietly starves curiosity. And we as SysAdmins let it happen.
I stood up Forgejo in my lab out of curiosity. It ended up replacing every GitHub component of my workflow. Zero regrets.
Lab work isn’t a hobby. It’s how you stay dangerous.
This week: do something that doesn’t have a ticket number.
From the Console
I attended a recent tech event in Texas which turned out to be a good event, had decent content, and a strong sense of community. At one point, I’m sitting there looking around the room….observing people as I’m want to do at times to gauge interest in a particular topic. I noticed lots of phones out, laptops open, heads down and my initial reaction was frustration, because I’ve been the guy up on stage talking, many times. Then I caught myself.
These aren’t checked-out people. These are DEPLETED people. The queue never stops. The alerts never stop. And somewhere in all that noise the thing that made most of us good at this job (the curiosity), the “I wonder what happens if I press this button” instinct, just quietly gets starved out.
That’s the thing worth talking about this week.
Folks that are regulars of the SysAdmin Weekly Podcast have likely heard me talk about the importance of curiosity for SysAdmins. I’ve been in the industry long enough now to know that SysAdmins used to have more time to be curious. We used to have more time for training. In fact, we’ve done our jobs so well over the last decade or two that tech and “digital transformation” have become core to many businesses. So much so, that the many businesses simply cease to function when tech systems are offline or impacted in some way.
Pair that with the fact that many organizations still view IT as a cost center. I still run into accounting heads that see IT as a black hole for money, and I still see leadership teams not giving IT teams the resources or the time needed to get the job done while leaving room for self-improvement or self-care. SysAdmins are overworked, and understaffed / underfunded, and it shows.
That all said, I would urge all of you to maintain that curiosity. Tinker in the lab. Check out that new community tool. Spin up that new game server. Most of us got into this industry because we have a deep love for the technology. Sometimes experimenting with new tech for no reason other than curiosity is all that’s needed to keep the bad parts of the job from overwhelming you and in the process it helps you keep your skills current and growing.
And now, back to our regularly scheduled programming.
The latest on the SysAdmin Weekly Podcast
Episode: 045 - Why Is It Always DNS? How DNS Actually Works, Breaks, and Gets Exploited Topic: Andy and Eric pull apart the full DNS resolution chain, Active Directory DNS integration, DNS security threats, and two war stories that will make you go check your fourth domain controller.
Why this one matters: - Most SysAdmins work with DNS every day without actually understanding how it works. This is the episode that fixes that - If you’ve ever said “have you tried flushing the DNS cache?” and quietly hoped it worked, this one is for you - DNS is also how attackers move data and talk to compromised machines. Understanding the protocol means understanding the threat
Episode Links
All Links:
https://www.sysadminweekly.com
YouTube:
Spotify:
The Take
I stood up Forgejo in my homelab recently. Self-hosted git forge, two-class repo model, nginx reverse proxy, internal HTTPS with mkcert. Did I plan for this to become my full git workflow? No. It started as curiosity. It ended up replacing every GitHub component of my day-to-day. GitHub no longer has a seat at the table, unless a repo has to be public and in that case I use repo mirroring so I still only have to interact with Forgejo at the end of the day.
On top of those benefits the curiosity that drove me to test out Forgejo to begin with enabled me to learn things I wouldn’t have learned any other way.
That’s the whole point. The lab isn’t always about the output. It’s about keeping the instinct alive. The instinct to dig in, break something on purpose, and figure out why it works. That muscle atrophies when you stop using it… and the environment most of us work in is perfectly designed to let it atrophy. Tickets, queues, SLAs, “what did you close today.”
It’s become a sad reality that the people who think lab work is a waste of time are usually measuring the wrong thing. They want deliverables. The lab doesn’t produce deliverables. It produces practitioners who don’t panic when something breaks outside the documentation.
This isn’t a generational thing either. I’ve seen 20-year veterans lose the curiosity instinct just as fast as someone two years in. It’s a culture problem. The fix is the same regardless of time in the chair: build something this week that nobody asked you to build. I’ll be money that two things will happen:
You’ll learn something new
You’ll enjoy the time spent doing so
So that said, what personal tech project are you interested in that you’ve been putting off? Maybe today is the day to start?
Community Signal
High-signal community work worth your attention.
mrlokans.work - “State of Homelab 2026” - A personal, opinionated stack writeup from someone who documents their gaps honestly: no backups, single NVMe, running costs under $10/month. The part worth reading is the SOPS with age encryption section as an upgrade over Ansible Vault for version-controlled secret management. Good real-talk energy, no vendor influence, highly recommend!
Brandon Lee - “How I Built a Self-Healing Home Lab That Fixes Itself” - Brandon Lee documents building an n8n automation pipeline that monitors CI/CD via Discord, uses an AI agent to identify failures, and automatically re-triggers failed pipelines without human intervention. Proxmox, Prometheus, Grafana, Docker restart policies. The interesting angle here isn’t the specific stack; it’s the idea that the lab itself can model the same automation discipline you’d want in production. Also a good read for this week!
Tool of the Week
Beszel - Lightweight infrastructure monitoring with a hub-and-agent model that actually fits the way homelabs and small shops are built.
Lightweight single binary, hub-and-agent model. Monitors CPU, memory, disk, network, and Docker/Podman container stats. The web dashboard is clean, alerting is configured with sliders and toggles instead of a rule language, and it doesn’t require standing up a full Prometheus stack to get meaningful visibility.
To be straight with you: this is a homelab and small fleet tool. No SNMP, no custom metrics, no HA, SQLite-backed. It is not a Zabbix or Prometheus replacement for enterprise environments with SLA requirements. For everything else, it’s the current community default for a reason.
Quick Win of the Week
Pick one service you interact with every day and actually read how it works this week. Not a tutorial. Not a YouTube walkthrough. The documentation or the RFC. Twenty-five minutes. The goal isn’t to become an expert; it’s to make contact with the actual thing instead of just operating the surface of it.
Fun Retro SysAdmin Fact
Before DNS existed, every host on the ARPANET was tracked in a single HOSTS.TXT file maintained by one person at the Stanford Research Institute and distributed via FTP. By 1983, the growth of the network had made manual updates impossible, which is what prompted Paul Mockapetris to design DNS. The specs he wrote in RFC 1034 and RFC 1035 are still the foundation of every DNS lookup happening right now.
Worth Your Time
Forgejo - User Documentation - If The Take in this week’s edition landed and you want to actually spin up your own git forge, the docs are legitimately good. The user docs index includes a getting started section.
Uptime Kuma - If you spin up Beszel, pair it with this. Beszel watches internal resource metrics; Uptime Kuma watches external service availability (HTTP, TCP, DNS, SSL). They cover different failure modes and the community runs them together for that reason. Worth mentioning that it also has a large number of GitHub stars….. meaning this one has been around long enough to trust.
Thing I’m Re-Thinking
The assumption: Staying current means following the news cycle.
Why it no longer holds up: Reading about technology and understanding technology are not the same thing. You can follow every CVE, every product announcement, every LinkedIn hot take… and still be completely lost the first time you have to operate something unfamiliar under pressure.
What I’m testing instead: One new thing in the lab per month. Doesn’t have to be big. Doesn’t have to be relevant to the day job. Just has to be something I haven’t run before.
Until Next Week
The queue will be there Monday. Your curiosity might not be if you keep telling it to wait.
Stay Frosty,
Andy
SysAdmin Weekly