Skip to main content

Gadget Wisdom

Category: Homelab

UPS battery backup protecting a home network closet with NAS, PoE switch, router and Ethernet equipment.”
0 Responses

Best UPS for a Home Network Closet: NAS, PoE, Cameras and Wi-Fi

Disclosure: As an Amazon Associate I earn from qualifying purchases. Product links may earn a commission at no extra cost to you.

A modem and Wi-Fi router are easy to protect with a UPS. A full home network closet is a different problem.

Once you add switches, wireless access points, PoE cameras, a NAS, a firewall, smart-home hubs and other equipment, you have more power draw, more devices that need to shut down cleanly and more reasons to care about what happens when the power stays out for longer than five minutes.

I use UPS units for both my networking gear and my NAS. My current NAS runs TrueNAS on a low-power Intel N100-based system, so it does not require a giant UPS by itself. But the NAS is only one part of the load. The router, switches, access points and other network equipment all need power too.

I have also had both sides of the UPS problem: an unsafe shutdown after losing power and a UPS battery that failed when I expected it to protect the equipment. Those experiences changed how I think about sizing a UPS. Capacity matters, but so do battery condition, monitoring, graceful shutdown and the reliability of the power where the equipment actually lives.

Quick answer: For a home network closet with a NAS, PoE switch and several other devices, a good starting point is looking at a 1500VA-class UPS with sine-wave output, USB or network monitoring and replaceable batteries. The CyberPower CP1500PFCLCD is the value choice I would consider first. For a NAS or server setup where management and expandability matter more, I would step up to the APC SMT1500C or Eaton 5P1500.

UPS Capacity Best For My Take Check Price
CyberPower CP1500PFCLCD 1500VA / 1000W Most home network closets Best value Check price
APC Back-UPS Pro BR1500MS2 1500VA / 900W Large home networks without server-class needs Good bridge between consumer and network-closet UPS units Check price
APC Smart-UPS SMT1500C 1440VA / 1000W NAS, servers and serious network closets My step-up APC choice Check price
Eaton 5P1500 1440VA / 1100W Premium network and server setups Excellent management-oriented option Check price
APC Smart-UPS SMT1500RM2UC 1440VA / 1000W Rack-mounted network closets Best fit if the rest of the equipment is already in a rack Check price

When a Router UPS Is No Longer Enough

I have a separate guide to the best UPS for a router and modem. If your entire critical load is an ONT or cable modem, router and perhaps one small switch, start there. You probably do not need the equipment in this article.

A network closet becomes a different problem when it contains things like:

  • A modem or fiber ONT
  • A router or dedicated firewall
  • One or more Ethernet switches
  • A PoE switch
  • Multiple Wi-Fi access points
  • Security cameras
  • An NVR
  • A NAS
  • A small home server
  • Home Assistant or other automation hardware
  • Smart-home bridges

At that point I care less about finding the smallest UPS that can technically support the equipment and more about building enough margin into the system that it remains useful as the network changes. Also, remember not every device needs to have redundant power.

Start With Watts, Not VA

Measure the actual network load instead of guessing using a smart plug or similar device. Knowing whether the closet uses 90 watts or 350 watts completely changes your considerations.

UPS marketing leads with VA ratings, but the watt rating is just as important. A UPS can be rated at 1500VA and still support a different maximum number of watts than another 1500VA model.

That is visible in the products above. The APC Back-UPS Pro BR1500MS2 is rated for 900 watts, the CyberPower CP1500PFCLCD and APC Smart-UPS models here are rated for 1000 watts, and the Eaton 5P1500 is rated for 1100 watts.

You do not want to add the wattage printed on every power adapter and assume that is your actual load. Those labels generally describe what the power supply can provide, not necessarily what the device consumes continuously.

Measure the whole network closet

If everything is already connected to one UPS, the UPS display may show current load. You can also use a plug-in power meter between the wall and the equipment while temporarily running it without the UPS.

Measure during normal operation and then consider what changes under load. A NAS may draw more power while its disks are active. A PoE switch may consume considerably more as additional cameras and access points draw power.

I would also leave room for the next device. Home networks have a habit of growing.

PoE Makes UPS Sizing More Interesting

A normal Ethernet switch has a relatively predictable load. A PoE switch is powering other equipment through the network cable, so the switch itself is only part of the calculation.

If the switch powers cameras, wireless access points, phones or other PoE devices, that downstream load is also being supplied by the UPS.

This is one reason backing up the switch can be so useful. One UPS connection may keep an entire group of cameras and access points online. It also means that adding another six PoE cameras later can materially change the amount of battery runtime you have.

When sizing the UPS, look at actual PoE consumption rather than simply the switch’s maximum advertised PoE budget. Then leave enough capacity for growth.

A NAS Changes What I Want From a UPS

A router reboot is annoying. An abrupt NAS shutdown is something I work harder to avoid.

A NAS may be writing data, updating filesystem metadata or running applications when utility power disappears. The goal of the UPS is therefore not necessarily to run the NAS for hours. It may be to keep the machine alive long enough for the outage to prove that it is not a momentary flicker and then shut the system down cleanly.

This is why I consider communications support essential when a NAS or server is part of the load.

TrueNAS and NUT

TrueNAS includes UPS support using Network UPS Tools, commonly called NUT. A compatible UPS can communicate with the NAS over USB, serial or, with suitable hardware, over the network.

TrueNAS can then react to UPS status and initiate a controlled shutdown rather than waiting for the battery to become completely exhausted.

The exact UPS and driver still matter. If automatic shutdown is important to you, check the current TrueNAS UPS documentation and the Network UPS Tools compatibility information before buying.

TrueNAS can also operate with one system acting as the UPS monitor while other systems receive status across the network. That is useful when several servers share one sufficiently large UPS.

If you are interested in the NAS side of my setup, I have an entire Building My NAS series covering the hardware and software decisions.

How Much Runtime Do You Actually Need?

This is where I think a lot of UPS recommendations go wrong. The correct size depends on the power where the equipment is located.

A network closet in a city apartment with an occasional one-second flicker has a very different problem from equipment sitting unattended at a vacation house where an outage might last six hours.

If your problem is brief flickers

Your main goal is continuity. You want the modem, router, switches and NAS to ignore a brief voltage drop that would otherwise reboot everything.

Runtime is secondary. Even a modest amount of battery capacity may be enough, although I still prefer reasonable headroom rather than running a UPS near its maximum load.

If you get occasional short outages

Now runtime becomes useful. If most outages last five or ten minutes, I want the network to stay up rather than immediately begin shutting things down.

A 1500VA-class unit becomes attractive because a network closet usually draws far less than the UPS’s maximum watt rating, allowing more runtime than it would provide to a heavily loaded desktop computer.

If outages are frequent

Battery wear and monitoring become more important. Every battery discharge uses part of the battery’s useful life.

This is also where Automatic Voltage Regulation is valuable. A line-interactive UPS with AVR can correct certain voltage variations without moving to battery every time the incoming voltage wanders outside its preferred range.

For a deeper look at AVR, waveform, communications and other UPS features, see How to Choose a UPS With the Right Features.

If long outages are common

A larger UPS may delay the shutdown, but eventually this stops being primarily a UPS-sizing problem.

If the power regularly disappears for hours, you may need a generator, a much larger battery system or another backup-power strategy. The UPS can bridge the gap, protect the equipment from short disturbances and give servers time to shut down properly.

If the network is at a remote or vacation property

This is the case where monitoring becomes especially valuable.

If nobody is there to hear the UPS beeping, I want some way to know that the property lost power and that the battery is being depleted.

There is another limitation: keeping your router alive does not guarantee that your internet service remains alive. Your cable company, fiber provider or other ISP has equipment upstream from your house. A neighborhood-wide outage may eventually take that infrastructure offline too.

If remote connectivity is important enough, consider cellular failover as a separate layer.

I discuss the rest of that problem in my guide to vacation-home remote monitoring.

Best Value: CyberPower CP1500PFCLCD

For many home network closets, the CyberPower CP1500PFCLCD is where I would start.

It provides 1500VA/1000W capacity, sine-wave output, Automatic Voltage Regulation, 12 outlets and USB communications. CyberPower also supports optional network management for the model.

That is a lot of capability for a home network, NAS and PoE setup without moving all the way into enterprise UPS pricing.

I would buy it for: a serious home network closet where value matters and USB shutdown is sufficient.

Check the CyberPower CP1500PFCLCD on Amazon

Best Consumer Step-Up: APC Back-UPS Pro BR1500MS2

The APC BR1500MS2 sits between the small consumer UPS and the more IT-oriented Smart-UPS line.

It is rated at 1500VA/900W, has 10 outlets, AVR, an LCD and a user-replaceable battery.

If your closet is mostly networking gear with perhaps a low-power NAS, this may be enough. If I were protecting a more substantial server or knew I wanted network-management options, I would move up to the Smart-UPS or Eaton choices below.

I would buy it for: a large home network that has outgrown a basic router UPS but does not need the management features of a higher-end unit.

Check the APC BR1500MS2 on Amazon

Best APC for a NAS or Server: APC Smart-UPS SMT1500C

The APC SMT1500C is where the product starts to feel more like network infrastructure.

It is a 1440VA/1000W line-interactive sine-wave UPS with eight outlets, AVR, a replaceable battery, USB/serial shutdown support, a SmartConnect Ethernet port and a SmartSlot for additional management options.

The difference is not simply another hundred watts. I am paying for a platform designed around servers, storage and network closets.

I would buy it for: a NAS or server I care about enough to justify spending more for management, monitoring and a more IT-oriented UPS platform.

Check the APC SMT1500C on Amazon

Premium Tower Choice: Eaton 5P1500

The Eaton 5P1500 is the premium tower option in this group.

It is rated at 1440VA/1100W, has eight outlets and true sine-wave output. Eaton includes USB and serial connectivity, with an optional network card for remote management.

The higher watt rating also gives a little more room for equipment growth, although runtime still depends on the actual load.

I would buy it for: a serious home lab or network closet where management, build quality and server-oriented features justify the higher cost.

Check the Eaton 5P1500 on Amazon

Best Rackmount Choice: APC SMT1500RM2UC

If the network already lives in a rack, the APC SMT1500RM2UC is the cleanest choice here.

It is a 2U rackmount Smart-UPS rated at 1440VA/1000W with six outlets, sine-wave output, AVR, SmartConnect, SmartSlot expansion and USB or serial shutdown support.

The advantage is organizational as much as electrical. Instead of putting a heavy tower on the floor or a shelf beside the rack, the UPS becomes part of the rack itself.

Remember that a rackmount UPS is heavy. Install it low in the rack and make sure the rack and mounting hardware are designed to support it.

I would buy it for: a permanent rack-mounted network installation where clean organization and IT-style management matter.

Check the APC SMT1500RM2UC on Amazon

Tower vs. Rackmount UPS

Choose a Tower UPS If… Choose a Rackmount UPS If…
Your network equipment sits on shelves You already have a standard equipment rack
You want the lowest cost for the capacity You want the UPS integrated neatly into the rack
You may move the UPS between locations The installation is relatively permanent
Rack space is limited Floor and shelf space are limited
You do not need rack-oriented cable management You want UPS and PDU wiring contained in the rack

Do not buy rackmount simply because it looks more professional. Tower units generally give a home user more flexibility and often better value.

Should You Use a Rack PDU?

A rack PDU can make sense when the UPS does not provide enough conveniently located receptacles for all the low-power equipment in a rack.

A proper PDU is a power-distribution device, not another UPS. Eaton specifically describes a UPS feeding a PDU as a normal basic network-closet architecture, and APC also distinguishes rack PDUs from ordinary surge strips.

Do not solve an outlet shortage by daisy-chaining UPS units. Do not casually add a household surge strip or extension cord to the output either. UPS manufacturers have specific requirements for what can be connected to their equipment.

If you need more outlets, check the UPS manufacturer’s guidance and use an appropriately rated PDU intended for the application.

Do You Need Pure Sine Wave?

For a simple modem and router, I would not make waveform the only buying criterion.

Once a NAS, server or other equipment with an active-PFC power supply is involved, I prefer a sine-wave UPS. It removes a compatibility question and is already standard on the CyberPower CP1500PFCLCD, APC Smart-UPS models and Eaton 5P1500 covered here.

This is one of those features where I would rather buy it once than diagnose a strange shutdown later.

Battery Replacement Matters More Than the UPS Brand Name

A UPS is not a permanent appliance. Its battery is a consumable.

I learned this the irritating way when equipment protected by a UPS went down even though there had not been an obvious outage. The UPS was still sitting there doing its job in theory. The battery was no longer capable of doing its job in practice.

I wrote about that experience in UPS Batteries May Only Last 2–3 Years — And They Usually Fail Without Warning.

For a network closet, I would:

  • Buy a UPS with replaceable batteries
  • Write the battery installation date on the UPS
  • Enable battery and self-test alerts
  • Check reported battery condition periodically
  • Test whether the UPS can actually carry the network load
  • Replace a weak battery before an outage exposes the problem

A UPS showing a green light is not the same thing as a UPS with a healthy battery.

What Should Be on Battery Backup?

Prioritize the equipment that either keeps connectivity alive or needs an orderly shutdown.

For my network closet, the battery-backed side would include:

  • ONT or modem
  • Router or firewall
  • Core network switch
  • PoE switch if it powers critical cameras or access points
  • NAS
  • Essential Wi-Fi access points
  • Home-automation controller
  • Critical monitoring gateways

I would be much less interested in using battery capacity for:

  • Monitors that do not need to stay on
  • Printers
  • Speakers
  • Decorative lighting
  • Chargers that can wait
  • Nonessential lab equipment

What Should Never Be Connected to a Normal Network UPS?

A UPS designed for computers and networking equipment should not become a general emergency outlet.

High-draw heating and motor loads can quickly overload it. I would not connect things such as:

  • Laser printers
  • Space heaters
  • Vacuum cleaners
  • Refrigerators or freezers unless the UPS was specifically selected for that load
  • Air conditioners
  • Large pumps
  • Other appliances with substantial startup current

Keep the UPS focused on the network and computing equipment it was purchased to protect.

Ancillary Gear I Would Buy With the UPS

The UPS is the major purchase, but a few inexpensive additions make the installation better.

A plug-in watt meter

Measure the actual network load instead of guessing. Knowing whether the closet uses 90 watts or 350 watts completely changes the runtime discussion.\

A proper rack PDU if you need more outlets

For rack installations, use suitable power distribution rather than a pile of consumer power strips. Follow the UPS manufacturer’s recommendations for connected distribution equipment.

An optional network-management card where it adds value

This is most useful when the UPS sits in a remote location, supports several machines or needs to be monitored without depending on one USB-connected computer.

The Eaton 5P supports an optional network card. APC’s Smart-UPS models include SmartConnect and a SmartSlot for additional management options. CyberPower supports optional remote network management on the CP1500PFCLCD.

A replacement-battery reminder

Put a calendar reminder in place when the UPS is installed. Waiting until the battery fails defeats a large part of the reason you bought the UPS.

My Buying Decision

If I were building a normal home network closet today, I would buy the CyberPower CP1500PFCLCD first. It gives me the combination of capacity, sine-wave output, communications and price that makes sense for most home setups.

If the NAS or server were more important, or if I expected to invest in better UPS management, I would move to the APC SMT1500C or Eaton 5P1500.

If the entire installation were rack-mounted, I would strongly consider spending the extra money on the APC SMT1500RM2UC rather than finding somewhere to put a tower.

But I would make that decision only after measuring the actual load.

A 1500VA UPS is not automatically better because the number is larger. The goal is to buy enough capacity and runtime for the equipment, give the NAS time to shut down cleanly, and make sure the UPS can tell you when its battery is no longer ready to do the job.

Frequently Asked Questions

What size UPS do I need for a home network closet?

Measure the actual wattage of the modem or ONT, router, switches, PoE devices, NAS and other equipment you want to protect. For a larger network closet containing a NAS and PoE equipment, a 1500VA-class UPS is a useful starting point, but the correct size depends on both total watts and the runtime you need.

Is a 1500VA UPS overkill for a router and modem?

Usually. If you are only protecting a modem or ONT and Wi-Fi router, see our router and modem UPS guide. The larger units in this article make more sense when a NAS, PoE switch, cameras, access points or servers are part of the load.

Should my NAS be connected to a UPS?

Yes, if the NAS contains data you care about. A UPS can protect it from brief outages and, with compatible communications and software, provide enough warning for a controlled shutdown during a longer outage.

Can TrueNAS shut itself down when the UPS battery is low?

Yes. TrueNAS uses Network UPS Tools for UPS monitoring and shutdown functions. Verify that your exact UPS is supported and configure and test the shutdown behavior before depending on it.

Should a PoE switch be on the UPS?

If the PoE switch powers equipment you want available during an outage, such as cameras or Wi-Fi access points, putting it on battery backup can protect all those devices at once. Remember that their combined PoE consumption is part of the UPS load.

How long will a UPS run my network?

Runtime depends on UPS model, battery condition and actual load. Do not estimate it from the VA rating alone. Measure the network’s wattage and use the manufacturer’s current runtime chart or calculator for the specific UPS.

Should I buy a rackmount or tower UPS?

Buy a tower UPS if you want flexibility and value. Buy rackmount when the network already lives in a rack and the cleaner installation is worth the added cost and rack space.

Will a UPS keep my internet working during a blackout?

It can keep your equipment powered, but it cannot guarantee that the ISP’s upstream infrastructure stays online. For a remote property where connectivity is critical, consider cellular failover in addition to UPS backup.

Can I plug a surge protector into a UPS?

Do not assume that an ordinary surge strip or extension cord is acceptable. Manufacturers including APC and CyberPower warn against common surge-strip configurations with their UPS products. If you need more rack outlets, use an appropriately rated PDU and follow the UPS manufacturer’s instructions.

How often should I replace a UPS battery?

Battery life varies with battery chemistry, temperature, discharge frequency and UPS design. Instead of relying only on age, monitor the UPS, perform tests and make sure the battery can still support the actual load. Replace a weak battery before you need it during a real outage.

Published on August 9, 2026
Full Post
0 Responses

Can’t Buy a New Router But You Can Build One

Recently, the FCC in the United States banned foreign made routers made by some of the most popular manufacturers…which means…well, pretty much all not already approved from popular brands. It means you can continue to buy current routers, but any new technology…not so much. The ruling is something of a joke. For one, manufacturers can just continue to sell old models of routers without new innovation.

The argument is that foreign entities can compromise these devices, however, they could compromise the existing ones, and even the new ones…they could deploy updates at any time and open a hole that wasn’t there previously. So, what is the solution? For me, that was build my own. I had gone back and forth on this over the years.

For one, if you wanted to stay with the most common routers, there exist many third party firmwares for consumer routers. These are open source options that give you more control, like dd-wrt, and openwrt. Routers are effectively little computers, and very limited ones at that. These routers have:

  • low power and relatively slow CPUs
  • extremely limited RAM
  • limited space to store additional programs

There is a lot of room therefore to move up. Instead of using a purpose-built router appliance, you can simply replace it with a low-power PC running router software Modern low power CPUs like the N100/N150 series are perfect for this function. You can use a mini PC, or there are manufacturers offering ones with specifications specifically in mind to act as firewall and router appliances. A more powerful processor gives your new router plenty of room for advanced networking functions.

To run my router, I chose Opnsense. Opnsense is an open source firewall and router, with a load of features and installable extensions.

  •  Enterprise-grade firewall controls
  • VPN support
  • traffic shaping
  • DNS filtering
  • detailed monitoring and logging

The default gives you a robust and secure router, and beyond that, any enhancement you can imagine can be added. And because it is modular, you can install additional functionality through plugins. This makes it far more flexible than most commercial routers.

What could be considered an advantage of building your own router is separating routing from Wi-Fi. It means you can upgrade and maintain these systems separately. These systems don’t actually have to live together. By addressing each piece individually, you can have better quality equipment, more reliability, and easier upgrades. If Wi-Fi standards change (and they always do), I can simply replace the access point without touching the router.

My Opnsense instance handles a lot more than a simple embedded appliance can. It:

  • Secure Site to Site VPN to multiple locations
  • Private VPN for my laptop and phone when on the go
  • Encryption for self-hosted services
  • Advanced firewall rules
  • DNS Management
  • Network isolation for IoT devices

For technically inclined users, it is relatively easy to get started and expand and learn over time. I cannot say how many times I have dug deep into the options when I added those VPNs, when I moved securing LetEncrypt certifications for my homelab onto the router, and more.

It isn’t just a long term solution, it is an adventure.

 

 

 

 

Published on April 6, 2026
Full Post
0 Responses

How to Choose a UPS: AVR, PFC, USB Shutdown, and Battery Backup Features

Buying a UPS sounds like it should be simple. Pick a battery backup, plug in the important stuff, and stop worrying about power outages.

Unfortunately, UPS shopping gets weird quickly. Some models are meant for routers and modems. Some are better for desktops. Some are designed for active PFC power supplies. Some can tell a NAS or home server to shut down cleanly. Some have replaceable batteries. Some are basically disposable power strips with a battery inside.

The right UPS depends less on the brand name and more on what you are trying to protect.

People buying power strips don’t think that much about them. But not only are UPSes designed to protect your devices, they are designed to keep them running during a power outage. How do you pick the right combination of features and price? In a previous post, I talked about how a problem with my UPS could have caused disaster.

Quick Answer: What UPS Features Actually Matter?

The most important UPS features are enough battery capacity for your load, automatic voltage regulation, replaceable batteries, surge protection, and a USB or network data connection if you need a computer, NAS, or home server to shut down cleanly. For modern desktop PCs and some servers, you may also want a UPS designed for active PFC power supplies.

Feature Why It Matters Who Needs It Most
Enough VA/watt capacity The UPS has to handle the devices plugged into it Everyone
Automatic Voltage Regulation Helps smooth brownouts and voltage dips without switching to battery Areas with unstable power
Replaceable battery Lets you replace the battery instead of throwing out the UPS Anyone keeping a UPS for years
USB/data port Allows clean shutdown for a NAS, desktop, or home server NAS, homelab, desktop PC users
Active PFC compatibility Helps avoid problems with modern PC/server power supplies Desktops, workstations, some servers
LCD/status display Shows load, runtime, voltage, and battery condition Useful, not always essential

APC vs. CyberPower: Which UPS Brand Should You Choose?

APC and CyberPower are the two consumer UPS brands most people run into first. I have used both. Neither is perfect, and I would not make the decision on brand alone.

For basic protection, either brand can be a big improvement over plugging important equipment straight into the wall. The better question is whether the specific model has the features your setup needs: enough capacity, replaceable battery, USB shutdown support, AVR, and the right waveform/PFC support for the devices you are protecting.

In recent years, I have mostly bought CyberPower units because the feature mix has worked well for my networking equipment and homelab gear. That does not mean every CyberPower model is the right model, or that APC is wrong. It means you should compare the actual unit, not just the logo.

AVR, PFC, and Cleaner Power: What These UPS Features Mean

The UPS feature list can look like alphabet soup, but a few features are worth understanding.

  • AVR, or Automatic Voltage Regulation: AVR helps correct voltage dips or surges without immediately switching to battery. That can matter if your power flickers, sags, or runs a little unstable.
  • PFC, or Power Factor Correction: Many modern computer power supplies use active PFC. If you are protecting a desktop PC, workstation, or server, make sure the UPS is compatible with that kind of load.
  • USB or data port: This lets a NAS, desktop, or home server know when the UPS is on battery so it can shut down safely before the battery dies.
  • Replaceable battery: UPS batteries wear out. If the battery cannot be replaced, the whole unit becomes a future e-waste project.
  • LCD/status display: Not essential, but useful for seeing load, runtime, battery condition, and voltage at a glance.

Choose the UPS Based on What You Are Protecting

A router, a NAS, and a desktop gaming PC do not need the same UPS.

Use Case What Matters Most What I’d Prioritize
Router and modem Long runtime at low power draw Efficient UPS, enough outlets, simple status monitoring
NAS Clean shutdown and uptime during short outages USB data connection, replaceable battery, enough runtime
Home server Graceful shutdown and stable power USB/network shutdown, AVR, PFC compatibility
Desktop PC Avoid sudden shutdowns Enough watt capacity, PFC compatibility, AVR
Security or smart-home gear Keeping monitoring online Runtime, router/modem backup, simple alerts

If your goal is mainly to keep the internet online during an outage, see my more specific guide to the best UPS for router and modem backup.

UPS Models I’d Consider

For a fuller-featured CyberPower unit, I would look at the CyberPower CP1000PFCLED or a similar model in that family. The reason to choose this tier is not just bigger battery capacity. It is the feature set: AVR, active PFC support, replaceable battery, data port, and LCD status display.

That kind of UPS makes more sense for a desktop, homelab server, NAS setup, or anything where a clean shutdown matters.

For lighter networking gear, the CyberPower EC650LCD can make more sense. It still has useful features like an LCD screen, replaceable battery, and data port, but it is better suited to lower-power equipment such as networking devices, small accessories, or a simpler monitoring setup.

I use the less expensive model for some of my networking equipment and the fuller-featured one for my homelab server. That split is the real lesson: do not buy one UPS model for every job just because it is familiar.

My Minimum UPS Requirements

For anything I expect to keep using for years, I want at least two things:

  • A replaceable battery: UPS batteries are consumables. If the battery cannot be replaced, the UPS has a built-in expiration date.
  • A data port: If the UPS is protecting a computer, NAS, or server, it should be able to tell that device when it is running on battery so the system can shut down cleanly.

For simple router/modem backup, the data port may matter less. For a NAS or home server, it matters a lot.

AVR and PFC support are not always mandatory, but they become much more important as the equipment gets more expensive or sensitive.

UPS Backup For Home Monitoring Systems

If you are using smart-home gear to monitor a second home, vacation home, cameras, leak sensors, or smoke/CO alerts, the router and modem become part of the safety system. If the network dies, the alerts may stop reaching you.

That does not mean every sensor needs a huge UPS. It does mean your modem, router, network switch, and possibly your camera/NVR setup deserve backup power.

For the broader monitoring setup, see my guide to vacation home remote monitoring.

Choosing the Right UPS

A UPS is not just a bigger power strip. It is part battery, part surge protector, part power conditioner, and sometimes part shutdown controller.

The right choice depends on what you are protecting. A router needs runtime. A NAS needs clean shutdown. A desktop needs enough capacity and PFC compatibility. A home monitoring setup needs the network to stay online long enough to send alerts.

Buy for the job, not just the brand.

Published on March 19, 2026
Full Post
UPS batteries fail silently
0 Responses

UPS Batteries May Only Last 2–3 Years (And They Usually Fail Without Warning)

Last week I walked into my office and discovered that several of my computer systems were down.

Everything about it looked like a power outage.

Except there hadn’t been one.

All of the equipment was connected to a UPS (Uninterruptible Power Supply) — essentially a battery backup power strip designed to keep systems running during outages.

But despite the name uninterruptible, everything had shut off.

UPS Batteries Don’t Last Forever

Most consumer UPS units use sealed lead-acid batteries. The typical lifespan is:

  • 2–3 years in normal conditions
  • shorter in warm environments
  • shorter if the UPS frequently switches to battery

Once the battery degrades enough, the UPS can no longer provide power and the unit may simply shut down when power fluctuates.

Many consumer UPS devices, such as mine,  provide little or no reliable warning before this happens. Silent failures defeat the purpose of a protective device.

My New Maintenance Plan

After this latest failure, I needed to adopt a much more systematic approach.

  1. Only buy UPSes with swappable batteries – the budget models often lack
  2. Label Every Battery – The date of installation of each battery will be labeled on the unit and on the battery.
  3. Perform Regular Hard Tests –

At regular intervals:

    • Power down connected equipment
    • Unplug the UPS from the wall
    • Confirm that the battery engages.
    • Optionally plug in a small device to verify the UPS actually provides power under load.

 

A UPS is not a “set it and forget it” device. If you rely on one to protect important systems — servers, networking gear, or workstations — you need to treat the battery as consumable component. Plan to replace it every few years and test it occasionally. Most UPS manufacturers recommend replacing sealed lead-acid batteries every 3 years, even if the device still appears to function normally.

Otherwise, the day you actually need the UPS may be the day you discover it stopped working long ago.

Published on March 12, 2026
Full Post
0 Responses

How to Supercharge Your Alerts with Ntfy: Priorities, Actions, and Smart Integrations

Previously, I wrote about using Ntfy as a notification system for Uptime Kuma. But the system has a lot of other advantages, so I wanted to go into a bit more detail on how it can be used.

Ntfy works on channels. Channels manifest as paths under the server URL. For example, ntfy.sh/gadgetwisdom would be a gadgetwisdom channel. By default, the ones on the public server are all public, meaning that you can only hide or restrict them by having a hard to guess channel name, if you host the server, you can add access control globally, or on a per channel basis to require authentication for subscribing, publishing, or both.

  • Priorities – Ntfy supports custom priorities from 1 to 5, with 3 being the default.
  • Templates – If a service sends data in JSON format, you can write a template for Ntfy to interpret it, rather than having to set up something in between to do that.
  • Action Buttons –  You can add buttons to view a webpage, send a http request to a URL, or on Android only, trigger an Android intent.
  • Attachments – You can send attachments using a file or URL link in a message. Files expire in a few hours, but linked files would be available longer.
  • Icons – To help identify different messages, notifications can have an icon.
  • Phone Calls – Ntfy supports via a Twilio account notifying people via phone using text to speech.

This leaves a lot of creative opportunities for integrations.

  • Maybe I want to have urgent messages phoned to certain people. Setting priority levels and noises can also help with this.
  • Action buttons could allow me to instantly react to alerts with simple actions.
  • Attachments already let me view snapshots from my Frigate feed, but what else could I send.

As I look to better integrate this, this service reveals a lot of simple options I can leverage.

Published on December 1, 2025
Full Post
0 Responses

Finding the Right File Synchronization Tool in Syncthing

Years ago, my first practical tool for file synchronization was a USB drive with a whopping 32 megabytes of storage. This was an issue for my mother, who, oddly for the time, just expected her files to follow her to every computer she went to. So I built a server to do just that. It was massive overkill for what it was actually used for.

Then came Dropbox. You installed an app and it synced everything in a designated folder to their servers. But Dropbox, like many services, became slower, bloated, etc.  So, I moved on to self hosting. First with Owncloud, and then with its fork, Nextcloud.  These platforms no longer bill themselves as file synchronization platforms, they bill themselves as ‘collaboration’ tools.

This is the problem that has taken me to look for a new solution that just does files. No more, no less. And while my instance of Nextcloud is still running, I am giving Syncthing a try. Syncthing requires no central server. It ticks all the security boxes: open protocol and encryption.

I installed a copy on my server, which allows me to synchronize the files there and back them up, in addition to syncing the various computers that need the files. It also, being peer to peer, allows me to have different folders synced to different systems, which allows for a bit more independent flexibility than the centralized systems, but also a bit more setup time.

Syncthing works perfectly on a local network, but has some challenges on the public internet. There are two services Syncthing provides for thi: A discovery server and a relay server.

Syncthing runs several public discovery servers, which assist in helping instances find themselves, but you can run your own. Discovery servers are safe enough to use publicly, although using them can identify that a device exists. The other server is a relay server. It allows relaying between devices that may be on different networks. The information is encrypted, so it is unlikely a malicious one could do much more than disrupt communication.

But regardless, I set up my own private discovery and relay server, and neither can be used without a token that I have to manually add to each device when setting them up. So, it is as secure as I can make it.

It remains to be seen if it will fully suit my needs, but so far, it is lightweight, flexible, and dumps all of the unnecessary cruft I do not want.

Published on November 24, 2025
Full Post
0 Responses

The Great Debate: One Big Computer or Several Smaller Computers

As part of my renovation project, as I make my new apartment into my own, and upgrade the infrastructure elsewhere in the building, I had several new lines run for cameras, all of which feed into Frigate, which I have mentioned previously. Frigate requires processing power, and I worry my server is taking on too many tasks that require the same resources.

The Problem

This brings me to my latest debate. Do I get one or more efficient computers with power efficient processors that can reliably handle low level tasks, or a single beefier computer? Technology has caught up to this while I was just letting my server run, because it just worked. Most home servers are idle most of the time, as you aren’t getting the demand a server with many users would receive, even with a full household. My NAS is running on an N100 CPU with 32GB of RAM.

The N100 has 4 cores, which is akin to having 4 different CPUs.  My server is running a processor that is a few generations older, with 6 cores. The benchmark of the N100 is about 1/3 of the score given to the older processor, but also uses a 10th of the power consumption. The NAS has an average usage of 3% CPU, as a NAS needs RAM to cache data, but serving it takes very little in terms of capacity.  By comparison, the server with the more powerful processor, running Jellyfin(a media server), Immich(photo server), Frigate(NVR), and more, is averaging 20% of more.

Plenty of CPU Maybe, but what about GPU?

Some of this involves hardware transcoding, using the GPU, to transcode video streams for Jellyfin. Frigate can use the GPU to run object detection. So, purchasing a Mini PC means another GPU, and the CPU can handle some low demand services. The alternative is to go for a an external GPU. However, the used computer my beefier CPU is in, is one I bought refurbished, only to discover that the PCI-E slot I would need to do that was not soldered onto the motherboard for some reason, which brought me to the idea of dividing the server, rather than replacing it entirely.

Multiple Computers…but in Software? Virtualization

Some people who have homelabs opt to go in a completely different direction. Virtualization. They have multiple virtual computers running on one server. There are advantages to this in that you can tear down a virtual computer and replace it, you can back up each piece separately. That leaves one with a lot of choices. I’ve read dozens of opinions and talked to several people about this.

Is more hardware better than virtualization? And if the idea of a low power CPU is to reduce energy costs, how does more than one of them help?

Final Thoughts and a question

As a final thought, having multiple pieces of hardware increases complexity, and thus maintenance time, but independent hardware means a failure in one system doesn’t mean a failure in another, the same isolation virtualization offers.

I have gone back and forth on a conclusion. I have seen people who run their entire infrastructure on a mini PC. That may be good for many people, but some might find themselves straining under the load.

But, what do you think? Am I better off upgrading or bifurcating?

Published on October 23, 2025
Full Post
2.5G vs 10G Ethernet home network comparison with switch, NAS, and network cables
0 Responses

2.5G vs 10G Ethernet: Which Home Network Upgrade Makes Sense?

Technology marches on, and keeping up with the times always raises the same question: When is an upgrade worth the cost? I remember when I first moved to gigabit networking. Prices have dropped since then, but it still surprises me how many devices in 2025 are stuck on older, slower Fast Ethernet. For homelab hobbyists, choosing anything below gigabit today would be unusual. Above that, the choices are: 2.5G, 5G, and 10G Ethernet.

Quick Answer: Should You Choose 2.5G or 10G Ethernet?

For most home networks, 2.5G Ethernet is the practical upgrade. It is cheaper, works over existing Cat 5e or Cat 6 wiring in many homes, and is fast enough for faster internet plans, Wi-Fi access points, and everyday file transfers.

10G Ethernet makes more sense when you have a NAS, homelab servers, large media files, video editing workflows, or multiple high-speed devices moving data across the local network. The best answer for many homes is not 2.5G everywhere or 10G everywhere, but a hybrid setup: 2.5G to most rooms and 10G links between the router, main switch, and NAS.

Upgrade Path Best For Main Tradeoff
2.5G Ethernet Most home users, faster internet, Wi-Fi access points, desktops Not as fast for NAS or server-heavy workflows
10G Ethernet NAS, homelab, large file transfers, video work, server links Higher cost, more heat, more careful hardware choices
Hybrid 2.5G + 10G Homes that want speed without rebuilding everything Requires planning switch uplinks and key devices

Future Proofing with the Right Cabling

As part of my recent renovations, I’ve been looking at ways to future-proof my infrastructure. The wiring I had installed is Cat 6, which has been the recommended standard for years and while there are newer options, Cat 6 supports higher data rates and distances, including 10G Ethernet up to 55 meters (180 feet)—more than enough for a residence.

But do you really need 10G? The cost-benefit question for home setups is still up for debate, though prices are slowly coming down.

The Internet vs. Local Speed

The first question is: what’s the speed of your internet connection? Even if it’s slower, there are still reasons to benefit from faster networking inside your home. For years, gigabit outpaced typical internet connections. But now, 1G and 2G internet plans are becoming more common—whether or not they actually deliver those speeds is another debate.

Why Faster Networking Still Matters

So why upgrade beyond gigabit? One reason is Network Attached Storage (NAS). I’ve written before about building a NAS. At the end of the day, a NAS is just a giant storage box—and when you’re pulling data from it, you want those files to arrive as fast as possible.

Serious homelab enthusiasts argue that now is the time to skip straight to *10G networking*. And there’s a solid case for that.

The Cost of 10G Hardware

A quick Amazon search shows this $33 PCI-E network card that uses a single SFP+ port. Dual SFP+ ports are only a few dollars more. I have an empty slot in my NAS, I could have this in there tomorrow. Switches with 10G ports used to be out of reach, but today I’m seeing options with 2.5G ports plus one or two SFP+ ports rated for 10G.

For those unfamiliar: SFP+ is a modular port that supports fiber optic or copper Ethernet modules, making it flexible for future upgrades.

A Hybrid Approach: 2.5G + 10G

My plan is to run a router with both SFP+ and 2.5G ports, with one of the 10G uplinks feeding a downstream 2.5G switch. That gives me:

  • A noticeable speed increase now
  • Flexibility to add faster devices later
  • The option to interconnect switches at 10G speeds

It’s not the full leap to 10G everywhere, but it’s a balanced, future-ready upgrade path for my homelab.

 

 

Published on October 9, 2025
Full Post
0 Responses

Planning for Homelab Failure: External Monitoring with Uptime Kuma + Ntfy

As much as I enjoy experimenting with smart home tech and running my own homelab, there comes a point where you just want things to work. After all, the fun of tinkering fades quickly if your light switches stop responding because of your latest experiment.

In a previous post, I wrote about building a B-Mode—the idea that your smart devices should still be functional, even in a degraded way, if your main controller goes down. But B-Mode only covers part of the problem. What happens if your entire homelab fails? What if your ISP drops, or your main server crashes while you’re away?

That’s where external monitoring comes in.

Why a VPS Is My Homelab’s Safety Net

I’ve always liked hosting as much as possible at home. It gives me control, keeps my data local, and aligns with the self-hosted philosophy I value. But there are limits:

  • If your ISP is down, your monitoring services hosted at home are offline too.
  • If your server crashes, you won’t know until you notice devices not working.
  • If you travel often, you may not realize something failed until much later.

The answer is to host a small part of your infrastructure outside your home. It is still self-hosting, even if you rent space. A cheap VPS (Virtual Private Server) can serve as a watchdog, constantly checking whether your homelab and internet connection are reachable. And it doesn’t take much to get started.

How I Use Uptime Kuma for External Monitoring

On my VPS, which costs just a few dollars a month, I run Uptime Kuma, a self-hosted monitoring tool that’s simple but powerful.

Uptime Kuma pings my home network and the servers of a few family members, keeping an eye on:

  • Internet availability
  • Critical self-hosted services
  • Response times

If a server or ISP connection goes down, Kuma alerts me right away. Because it’s running on a VPS outside my home, I’ll still get notified even if my entire local network is offline.

For anyone running a homelab, Uptime Kuma is one of the easiest and most effective ways to add external monitoring.

Getting Alerts Anywhere with Ntfy

Monitoring isn’t enough unless you also get alerts you’ll actually notice. That’s where Ntfy (pronounced notify) comes in.

Ntfy is a self-hosted notification service. It can be installed on your VPS, or you can use the hosted version for free with optional paid upgrades to support development. Ntfy supports a web app, desktop, and mobile apps—and it integrates smoothly with Uptime Kuma and other services.

Here’s how I use it:

  • Uptime Kuma sends alerts to Ntfy when a monitored service fails
  • Ntfy pushes those alerts to my phone and desktop instantly
  • Other tools (like Frigate for camera detection) also send notifications into Ntfy

Ntfy organizes messages into topics that you subscribe to. With access control, you can decide which users can read, write, or subscribe to specific topics. It’s lightweight, flexible, and can be wired into nearly any self-hosted project.

Designing for Outages You Can’t Control

With this setup, I’m covered for most common homelab issues:

  • Server outages
  • ISP failures
  • Device-specific crashes

The only remaining weak point is the VPS provider itself. But outages at large hosting companies are rare, and if you’re truly paranoid, you could run monitoring on two different VPS providers for redundancy.

Even then, my monitoring chain is still stronger than it would be if everything lived only inside my home.

Why Homelab Reliability Matters More Than Perfection

Experimentation is fun, but reliability matters. By running Uptime Kuma and Ntfy on a low-cost VPS, I’ve given myself peace of mind that if my homelab or ISP ever fails, I’ll know right away.

Just like my B-Mode philosophy for smart devices, this approach ensures I’m not left in the dark when things go wrong.

Because in the end, the smartest homelab isn’t the one with the most toys—it’s the one that keeps working even when something breaks.

Published on September 10, 2025
Full Post
Illustration of a person monitoring POE security cameras using Frigate NVR software on a computer, with outdoor cameras mounted on a house and detection alerts shown on screen.
0 Responses

POE Cameras and Frigate NVR: Why I Switched to Local Home Surveillance

During my recent renovation, I added two additional cameras to my new space, at the two points of ingress. This was something of a departure as these were also the first Power Over Ethernet(POE) cameras I’ve had installed, as I had someone on-site available who could run the cables cleanly.

I’ve tried a variety of ecosystems for cameras, both for myself and others. Many of them push you toward subscription-based cloud services, which features like video history, motion detection, and notifications only work fully if you pay monthly. Some of them barely provide any features without paying, despite the fact you bought the device.  Even when offering local options, this is often storage with a microSD card in the camera, which is clunky, slow, and unreliable.

That is why I decided to go with a network video recorder. A server that takes the feeds from all the cameras and stores the recordings. You can buy commercial NVRs you can purchase and install in your house, including some that integrate with the specific hardware cameras you bought, but I wanted a solution that aligned with my philosophy of self-hosted, privacy first smart home tech.

So I chose Frigate.

Quick Answer: Why Use POE Cameras With Frigate?

POE cameras and Frigate make sense if you want a local home-surveillance system that does not depend on a cloud subscription. Power Over Ethernet cameras are more reliable than Wi-Fi cameras, and Frigate can record video, detect people, cars, animals, and other objects, and let you tune alerts around the parts of your property you actually care about.

Choice Why It Matters
POE cameras More reliable than Wi-Fi cameras and powered through the network cable
Frigate NVR Local recording and object detection without relying on a cloud camera plan
Detection zones Reduce false alerts from sidewalks, streets, neighbors, or passing cars
Self-hosted storage Keeps recordings under your control instead of inside a camera company’s subscription
Frigate+ Optional model improvements without turning the whole system into a traditional cloud lock-in product

Why Frigate?

Frigate is an open-source NVR designed for real-time object detection all running on local hardware. It is deeply customizable and can be tuned to only record what matters to you – people, cars, or animals, depending on what zones and filters you decide.

For example, one of my outdoor cameras flagged every pedestrian across the street, which is well outside of the zone I am concerned about. I can narrow the zone to only my property, to dramatically reduce noise in footage and alerts.

Frigate recently added:

  • facial recognition
  • license plate recognition.
  • View-only user roles for shared access

Everything is processed locally, with no cloud dependency.

Frigate+: Smarter Detection, Optional Subscription

To improve detection, you can also subscribe to Frigate+, a $50/year subscription which offers better trained models for detection. These are trained by other users of Frigate. You can participate by submitting false positives and other information voluntarily. If you cancel, you get to keep the downloaded models, you just stop getting updates.

This helps support the developers and doesn’t lock you into a traditional subscription model.

Frigate Notifications

One gap in the core Frigate setup is the lack of built-in robust multi-platform notifications. That’s where another piece of software, Frigate-Notify, comes in. It offers all of the notification options I might want.

  • Rich notifications
  • Cross-platform delivery including mobile, desktop, and messaging apps
  • Fully customizable

Next Steps For My Frigate NVR

Inspired by how well the new system is performing, I plan to replace more of my older Wi-Fi cameras with wired POE models for improved reliability. Wired cameras streaming directly to my NVR reduces lag, improves reliability, and gives me full control over recording, storage, and alerts—without the cloud.

If you’re tired of cloud lock-in and unreliable Wi-Fi cams, and you want a privacy-respecting, smarter surveillance system, Frigate + POE may be the combo you’ve been looking for.

 

Published on September 8, 2025
Full Post

Get New Posts By Email