What should I invest time learning these days?

Posted by ss2014s@reddit | sysadmin | View on Reddit | 36 comments

I'm a middle-aged sysadmin and want to keep growing and not become stagnant. What would you all say are some worthwhile technologies/topics to invest time into learning?

Ideally, I'm hoping to learn something that's both useful in the IT job market (future-minded) and is also fun/interesting to me. That'd be great to check both boxes if possible.

A short list I came up with so far that sound interesting to me (unsure how many of these are useful in the future IT job market):

Docker/containers -- I have experience with VMs but next to none with containers
Learning about how AI actually works
OpenClaw -- maybe set it up in a container and play with it (carefully)
Running local AI models in Ollama/other platforms and messing around. I only have a RTX 4070 Mobile GPU w/ 8GB of RAM to mess around with, but hey it's better than nothing.
TryHackMe/HackTheBox learning path -- I have some cyber sec experience but could also learn more/get hands on and refresh my knowledge
Cryptocurrency -- I have zero hands-on knowledge. Seems like it'd be a good thing to know more about, ie: how do you pay someone in crypto, etc...
Arduino/Raspberry Pi/etc -- I know nothing about microprocessors or basic electrical circuits, etc.
Modern web application technologies/tokens/code -- again, zero knowledge here.

I'm open to other ideas, please!

I'm comfortable around a CLI, PowerShell, common networking protocols, Linux, OpenWRT, firewalls, Citrix, VirtualBox, etc.

Thanks!

[-]

jreykdal@reddit

Docker is fun.

Recently started myself and find it useful.

[-]

Speaking with IT MSPs, most of the customers are all in on implementing AI in something. If you want to be valuable (and future-proof), become the expert on implementation (with guardrails) and support.

Don't waste your time on crypto unless you have a financial interest. Running local models takes maybe an hour or so of your time before you quickly realize that you need strong hardware to do anything useful. OpenClaw was fun.

[-]

hihcadore@reddit

How to hang drywall or raise sheep

[-]

Wolfram_And_Hart@reddit

Wood working

[-]

hihcadore@reddit

Exactly! That’s what I do!

I even decided to go the hand tool route so I don’t chop my fingers off. With AI I might forgo that since you can talk to it, and it can do 90% of what I’m getting paid for.

[-]

MathmoKiwi@reddit

Or goats?

[-]

Curious201@reddit

you have a decent admin foundation already, so i would not jump into five new areas at once. if your goal is to stay useful and not get boxed into old-school support, i would put most of the time into automation, cloud identity, and containers, but in a practical way. learn enough Python to write small tools around APIs, log parsing, reports, and boring cleanup jobs, not abstract leetcode stuff. learn Docker well enough to deploy and troubleshoot small services, because even if you do not become a full DevOps engineer, containers show up everywhere now. for cloud, i would start with identity, networking, storage, and backups in Azure or AWS instead of chasing every service. the hardware/ESP32/CNC stuff is fun and can absolutely be useful, but for sysadmin career value i would treat it as a side track unless you are moving toward industrial/OT work.

[-]

shiranugahotoke@reddit

Containers! I think this keyholes into the modern web applications as well.

If you have a strong understanding of architecture and infrastructure, you can get pretty dang far with a coding agent - even if you don't have strong coding skills.

I built a phone provisioning server to provide XML configs to our Cisco phones over HTTPS, upgrade firmware, and track phone assignment state against extensions in our RingCentral tenant.

It uses the API to get the user / extension info, build the config file, and serve it to the phone.
It's all contained in a single docker container that uses \~100MB of memory.
It's built on all update to date and pinned package versions of react, python, and fastapi, stores any credentials encrypted in the DB with a runtime key with optional key vault integration, certificate management, SSO integration, and has a super fast UI.
The container rebuilds in about 30 seconds and boots in about 10.

It took me 3 days to get the project fully working and another probably 3 days to tweak and refine and add features I forgot.
I've run it against multiple static code analysis tools and performed a 'Security Audit' and so far it's looking pretty good - better than most of our legacy apps.

[-]

ss2014s@reddit (OP)

Lots of votes for containers so far! Sounds like you built a really neat tool, with a container as the foundation. Nice work.

[-]

ElectricOne55@reddit

Is Terraform used a lot? I see it on a lot of applications, but idk if it's worth the time or effort to learn? Or to just stick with cloud platform paths?

[-]

uptimefordays@reddit

Over the past five years, I’ve extensively used Terraform for provisioning both on-premises and cloud infrastructure. Pulumi is also gaining popularity, but it has a steeper learning curve. Additionally, overcoming the institutional inertia of existing tooling is challenging. Many companies already have Terraform and are reluctant to switch to something else for developer and engineering experience.

[-]

Logical_Sort_3742@reddit

If you can handle containers, the road is short to pods. If you can handle those, the road to Kubernetes is short. Not easy, but short. It all naturally coheres. So you can build a lot of highly relevant skills on top of eachother.

[-]

MalletNGrease@reddit

How to write a good resume.

[-]

AddendumWorking9756@reddit

Containers and blue team work win, the other two have weaker ROI for someone already deep in IT. Skip the guided platforms once basics are there, real-artifact stuff like CyberDefenders gets you further than walkthroughs.

[-]

1stUserEver@reddit

How to fix things and be frugal. Times be tight.

[-]

mgcjr1@reddit

ITIL 4 Foundations is another good option for shoring up Service relations and operations.

[-]

SurfaceHub2S@reddit

A lot of expected IT solution focused answers here, but here’s my take considering that your goal sounds like it's probably increasing your earning potential.

If you can’t communicate your impact, you’re essentially invisible. Do yourself a favour and master tools like Power BI, PowerPoint, even just how to write an effective email to improve data storytelling.

Stop being the person who just "presses buttons" and become the one who moves the needle. For example: if you decommission five VMs without a paper trail, nobody cares. But if you frame it as a solution to a resource bottleneck and show the $$$ saved, you’ve turned a routine task into a strategic win.

If you don't document the value, it didn't happen and you will struggle to justify a promotion.

[-]

shimoheihei2@reddit

I don't really think the answer changed since years ago. You should learn how to do proper troubleshooting, solve complex problems, do efficient web searches and communicate properly. Any specific technology is going to change constantly. What you learn now will likely be obsolete in 5 years.

[-]

hyguru6@reddit

Im not sysadmin but I learn the field. I heard opinion that AI will have to run on something so Docker>Kubernetes>OpenShift.

[-]

sembee2@reddit

Goat farming. The earlier you start the more informed you will be. Also look at llamas and alpacas, as they can be worth more, although the startup costs are higher as you need bigger fences.

[-]

Hairy-Link-8615@reddit

So in a similar position as you.

Imo don't drive into Cryto too much cis it might suck you in. 😂 It's deep once you start looking at trading and meta etc

Pie is fun, retro game something

But 100% anything cloud is good. If you want a recommendation.

Orcole cloud offer a free lifetime tier cloud box.

Go install something with swag docker. That will hit a bunch. Then you can do the same on the 1 year tiers in azure and aws. Buy a cheap domain to cover cert fees.

3.clouds docker cert renewal with let encrypt.

Then hopefully you'll have something that takes your fancy.

My little vps took me an afternoon to fiddle with and the services

[-]

Sree_SecureSlate@reddit

Diving into Docker and local AI with Ollama may be a total blast.

It's the perfect way to turn your existing CLI skills into modern DevOps and "AI-Ops" without feeling like you're starting from scratch.

[-]

KiLoYounited@reddit

I’d say look at deploying and orchestrating container services.

Along the lines of coding, consider making tools for you and the other sys admins. Nothing really beats a good home grown tool that fills a niche or role your team needs. I’m assuming you have basic PS or bash skills, Python is not a massive leap. There is a stigma from “real developers” regarding Python, but ignore it. Neither you or I are building apps that have large user bases or require the features that C/Rust/whatever have.

[-]

Opposite_Bag_7434@reddit

Docket, containers, anything Infosec, local and enterprise cloud AI of every flavor, strong familiarity of the major cloud platforms as well.

Every enterprise is going to be a bit different so you almost can’t go wrong.

[-]

Amomynou5@reddit

I would say look at setting up your own homelab first, because once you do that, you'll organically start gaining skills in multiple areas, and it's way more fun this way instead of trying to force yourself to learn some particular technology that you can't relate to.

Start by getting some cheap second-hand / ex-lease PCs (small form-factor PCs like the HP EliteMini/Dell OptiPlex/Lenovo ThinkCentre etc). Wire them up using the fastest network you can afford, or just go with what you've got right now. Install Proxmox and Kubernetes. Maybe even dedicate another machine to be used as a NAS. Look at setting up a firewall VM like OPNsense, and maybe some IAM solution like Authentik. Cloudflare or Traefik for external access / reverse proxy / SSL.

Then look at hosting some real-world apps that you might actually use, like Immich for photo/videos, ownCloud OCIS for file hosting, home assistant for home automation, AdGuard Home for adblocking and so on. You'll learn way more and have much more fun doing so.

[-]

vogelke@reddit

I'd go with Docker/containers, and then set up your own webserver (Apache, Nginx, whatever). A local webserver is incredibly useful.

[-]

Opposite_Bag_7434@reddit

This for sure

[-]

ss2014s@reddit (OP)

Thanks, great idea!

[-]

toyatsu@reddit

Containers, Hacking imo these are the things that won't die in the next few years.

Arduino and raspberry Pi are also nice, but more for private use than in a company, but youre gonna learn a lot

[-]

Lonely-Candidate-231@reddit

geez i’m help desk and know all of those topics pretty well. i guess just get into studying at home and building a homelab with automated pipelines similar to enterprise environments

[-]

CherrySnuggle13@reddit

If you want future-proof + practical, I’d lean into containers (Docker/K8s), cloud fundamentals, and identity/security (IAM, zero trust). Add some scripting/automation around that. Homelabs help a ton. The fun stuff like Pi or local AI is great too, but the first ones pay off faster career-wise.

[-]

ss2014s@reddit (OP)

Cool, thanks! Sounds like solid advice 😄

[-]

Draculalol@reddit

Kubernetes is black magic, maybe start there

[-]

ss2014s@reddit (OP)

I should probably get a bit more familiar with containers first though, right?