Your Alarm Still Beeps. That Does Not Mean Anyone Is Being Told.

Somebody changes the internet. A new provider, an upgraded modem, a mesh system going in, or just a Wi-Fi password finally being changed to something that is not written on the side of the old box. It takes an afternoon, everything important reconnects, and life goes on.

Six months later the alarm has still never once failed to arm. The keypad beeps on the way out of the door. The little green light does what it always did. And there is a reasonable chance that for those six months nobody at the monitoring centre has heard a word from it.

This is worth separating out carefully, because the two halves of a monitored alarm fail independently and only one of them makes a noise. A local alarm makes noise. A monitored alarm makes noise *and* tells somebody. Arming the system every night tests the first half thoroughly and the second half not at all. We are a computer and home-networking shop rather than an alarm company, which turns out to be the right trade for the half of this that actually breaks — because it is nearly always a network problem, and it nearly always started on the day the router changed.

The failure that has no symptom

Almost everything an alarm panel does happens inside your house and needs nothing beyond the panel itself. Arming and disarming, the entry delay, the sensors reporting a door opening, the keypad tones, the siren: all of it is local. Cut the panel off from the internet entirely and every one of those still works exactly as it did before.

The one thing that leaves the building is the report — the message that says a door on this address just opened while the system was armed, sent to a monitoring centre where a human being decides whether to call you and whether to dispatch. That path is the thing you are paying a monthly fee for, and it is the only part with no local symptom when it breaks.

The manufacturers are perfectly clear about the stakes when the path is down, in the parts of their documentation almost nobody reads. Alarm.com, whose platform sits behind a very large number of professionally installed systems in this country, describes its "Radio Not Responding" trouble condition in flat terms: your security panel "is not able to communicate with the network" and "may not trigger an alarm or signal for help in an emergency." The instruction attached to it is a single line — contact your service provider.

Compare that with how the rest of your technology behaves. A laptop that loses Wi-Fi puts a symbol in the corner and stops loading pages. A phone tells you immediately. A printer sulks visibly. An alarm panel that cannot reach its monitoring centre carries on being a perfectly good alarm panel, and the gap is discovered on the one night it matters.

It is almost always the router, and almost always months ago

When we get called about this, the cause is boringly consistent and lives entirely on our side of the fence.

The Wi-Fi password changed. A panel or its communicator holds the password that an installer typed in once, on the day it was commissioned, possibly years ago. There is nobody sitting at it to re-enter anything. Change the password and it is locked out permanently, still trying.

The router was replaced. New provider, new modem, a mesh kit going in over the top. If the network name is different — and provider-supplied gear arrives with its own default name printed on a sticker — the panel is looking for a network that no longer exists.

The bands got merged. Nearly every router sold in the last few years broadcasts 2.4 GHz and 5 GHz under one network name and decides for each device which to use. That is excellent for laptops and poor for stubborn 2.4 GHz-only equipment, which is what a great deal of alarm and sensor hardware is.

Ring says the quiet part out loud in its own troubleshooting for the Alarm base station: "If there have been any updates to your wifi network, password, equipment, or internet provider, confirm that network settings are accurate." That is the whole diagnosis in one sentence, and it is the sentence nobody reads until after the fact.

The reason this goes undetected for so long is the gap between cause and discovery. The router changed in March; the question "is the alarm still reporting?" occurs to somebody in October, if it occurs at all. Nothing links the two events, because nothing ever announced itself. It is the same shape as a solar gateway that stops reporting or a printer that vanishes off the network after a router swap — with the difference that when the printer stops, you find out that afternoon.

Offline is not the same as unmonitored — find out which one you have

Before assuming the worst, establish which kind of system you own, because the consequences of a dead Wi-Fi connection are wildly different depending on the answer.

Many monitored systems are dual path: they report over your internet connection when it is available and over a cellular radio built into the panel when it is not. If yours is one of those, losing the Wi-Fi may cost you nothing more than the app on your phone, while the alarm reporting carries on over a network you do not own and cannot break. If yours is single path over your home internet, then your router is the alarm's only route to the outside world, and everything on this page applies with force.

Consumer kit works the same way, with the subscription attached. Ring's own documentation on a base station going offline is a good worked example: the system "will continue to monitor your sensors and send alarm signals," the device "remain[s] armed," and "the Ring Alarm siren will still function" — but you will be "unable to access or control your devices through the app." The cellular fallback that keeps the reporting alive, Ring states plainly, "requires a compatible Ring subscription." Its battery arrangement is documented in the same place: around fifteen minutes of normal operation, then up to twenty-four hours in a reduced-power mode that keeps sensors and alarm signals going.

So there are three genuinely different states, and people routinely confuse them: the app is down but monitoring is fine; the app is fine but monitoring is down; or both are down. Do not infer any of them from the panel. Ask.

The question to put to whoever monitors your system is short and specific: "Is this panel dual path, and if so, which path did my last signal arrive on?" If the answer to the first half is no, the follow-up is what it would cost to add a cellular communicator — most panels of any age can take one.

The signal your system sends when nothing is happening

Here is the mechanism that makes this checkable, and it is the single most useful thing on this page.

A monitored alarm does not only transmit when something happens. On a schedule, it sends a message whose entire content is "I am still here and I can still reach you" — the industry calls it the periodic test, the timer test, or the supervision signal, and typical practice is one every twenty-four hours, sometimes stretched to every couple of days on older or lower-tier accounts. If the monitoring centre stops receiving them, that shows up on their side as a late or missing test. They know your alarm has gone quiet, and they usually know within a day or two.

Which raises the obvious question: if they know, why did nobody call? Sometimes they do. Sometimes the notification goes to the dealer who installed the system rather than to you, and that dealer has since been acquired, renamed, or has gone out of business. Sometimes it goes to a phone number or email address that was accurate in 2016. And sometimes the contract simply does not commit anybody to telling you. None of that is universal and we are not going to pretend it is — the point is that you cannot assume silence means everything is fine, because silence is exactly what a dead reporting path produces.

The good news is that this makes a genuine health check available for the price of one phone call and no equipment at all. Ring your monitoring company and ask: "When did you last receive a test signal from my system, and what path did it come in on?" A company that monitors your alarm can answer that in under a minute. If the last one was four months ago, you have your answer, and you did not have to touch anything to get it.

Do that call first. Everything else below is what you do when the answer is unsatisfying, or when you want to confirm the whole chain end to end.

How to test an alarm properly — and the phone call almost everyone skips

Testing a monitored alarm is not difficult, but doing it without warning anybody is how people end up with a police cruiser in the driveway and an invoice a few weeks later. The order matters.

One: put the system on test with the monitoring centre. Call them, identify yourself with the account details and the verbal password on the account, and ask for the system to be placed in test mode for a set period. Some platforms let you do it yourself — Alarm.com exposes it in the app under Menu, then Monitoring Settings, then System Test Mode, in one or two-hour blocks, though only "the primary login and secondary logins with master permissions can put the system on test mode." Alarm.com describes what test mode actually does without euphemism: "The monitoring station is not going to respond to alarms (i.e., dispatch police) from the system during that time." When the window elapses, "the account will automatically exit test mode."

Two — and this is the step that is missing from essentially every national article on the subject — notify your city police department as well. Putting the system on test stops the monitoring centre dispatching, but your local ordinance is a separate rulebook with its own definition of what counts as a false alarm, and in Southern California those ordinances tend to require prior notice to the police directly. Glendale's alarm ordinance is a clean example: its municipal code excludes from the definition of a false alarm anything "resulting from the authorized servicing, testing, maintenance, adjustments, alterations or installations of such alarm system" — but only "provided that the respective permittee or authorized agent thereof, notifies the police division emergency communications supervisor prior to commencement of any of the aforesaid activities." Prior notice is the whole exemption. Without it, your test is a false alarm like any other.

Three: run it. Arm the system the way you normally would, wait out the exit delay, then trip it deliberately — open a protected door, or walk a motion detector. Let the siren run for a good thirty to forty-five seconds so the panel definitely transmits, then disarm with your code.

Four: call back and ask exactly what arrived. Not "did you get it" — ask which signals were received, at what times, and over which path. If you have smoke or carbon monoxide detectors on the system, test at least one of those too, because they can report to a different destination than the burglary side and can fail independently of it. Ask about the periodic test signal while you are on the line.

Five: take the system off test, and confirm out loud that it has been taken off test. A system left in test mode is exactly as unmonitored as one with a dead router, and this is the most common way a well-intentioned test makes things worse.

Do this after any change to your internet or Wi-Fi, after any electrical work or power outage that took the panel down, when you move into a house with an existing system, and otherwise about once a year. It takes fifteen minutes and it is the only thing that tests the half of the system you cannot see.

The part that costs money: your city already has an opinion about this

Alarm permits and false-alarm charges are a local matter, and every city we work in handles them slightly differently. The pattern is remarkably consistent though: you need a permit, you get a small number of free police responses, and after that the charges escalate — and testing only escapes the charge if you gave notice first.

Pasadena requires a permit for an alarm system at a residence or a place of business. On the fee schedule effective July 1, 2026, the permit for operation of an alarm system is $39.54 and runs for a year, prorated from the application date. The first false alarm in the annual billing cycle, which runs January to December, carries no charge; the second and each one after it is $181.00. Operating an alarm system without a permit at all is a separate $163.00 charge, which is levied on top when a response turns up at an unpermitted address.

Riverside puts the requirement in the ordinance itself: "No person shall install or cause to be installed, use, or maintain an alarm system on premises owned, rented or leased within the City without first having obtained an alarm user's permit." The permit is a one-time $16.00 application fee. For a permitted location the first two responses are free, the third is $100, the fourth $150, the fifth $200, and the sixth and beyond $250 each. Unpermitted locations are charged considerably more, and from an earlier point in the ladder.

Two practical consequences follow. The first is that a system you test carelessly twice in a year can cost you more than the monitoring did. The second is quieter and catches people out when they move: the permit attaches to the address and the user, so buying a house with an alarm already on the wall, or renting one, does not inherit you a permit. Check.

Fee schedules change every year or two, so treat the figures above as illustrations of the shape rather than as current everywhere — look up your own city's alarm page before you test. What does not change is the underlying rule, which is that the police want a phone call before you set your own alarm off on purpose.

The network side, briefly, because we have written it out elsewhere

The fixes that keep a panel, a communicator or any other awkward device attached to your network are the same ones that keep a solar gateway, a wireless printer and a smart thermostat attached, and we have covered them at length on other pages rather than repeat them here. In short:

When you replace a router, give the new one the exact same network name and the exact same password as the old one. Every silent device in the house then reconnects on its own and you never find out how many of them there were. This is ten seconds of work at setup time and it prevents the entire problem.

If the password genuinely has to change, write the list of silent devices before you change it — alarm panel, cameras, doorbell, thermostat, solar gateway, garage opener, printer, robot vacuum — because none of them will tell you they are locked out.

Give the 2.4 GHz band its own network name, or use the dedicated IoT network many routers now offer, and put the awkward gear there. Keep the name plain: letters, numbers and hyphens, no spaces, apostrophes or emoji, because embedded hardware handles those badly.

And if the panel or its communicator has an Ethernet port and there is any reasonable way to reach it with a cable, use it. A wired connection removes every problem in this section at once, and if there is no network jack where the panel lives, existing coax and MoCA will usually get you one without opening a wall.

A camera is not an alarm, and a doorbell is not either

Worth stating plainly because the marketing has thoroughly blurred it. A camera that goes offline stops recording and stops notifying your phone, which is annoying and occasionally expensive. A monitored alarm that goes offline stops anyone being dispatched, which is a different category of problem. They fail for identical network reasons and they do not carry identical consequences.

Modern kit muddies this further, because the same manufacturer often sells you a doorbell, some cameras and a monitored alarm on overlapping subscriptions, and it is genuinely easy to believe that a plan covering one covers all of them. Find out which of your devices is professionally monitored, which is self-monitored — meaning the only person being notified is you, via an app, on a phone that may be face-down in another room — and which is neither. Then decide whether that matches what you thought you were buying.

Whose account is it, and what happens when nobody is paying

A large share of residential alarm systems were set up under the installing dealer's account with the homeowner given a limited view, which works fine right up until the dealer stops existing. If the app has stopped loading, the company's number rings out, or the name on the paperwork does not match the name on the invoice any more, that is the first thing to untangle — and it is a business problem rather than a technical one.

The other common case, and there is nothing wrong with it as long as it is deliberate, is a panel that nobody is paying for. Monitoring was cancelled at some point, the equipment stayed on the wall, and it still arms and still sounds. That is a local alarm: a deterrent and a noise, with no one at the other end. Perfectly legitimate, and a real problem only when somebody in the household believes it is still monitored. If you are not sure which you have, the last-test-signal phone call above settles it immediately, because there is nobody to phone.

If you do want monitoring back on an existing panel, ask before you assume: whether that specific panel can be taken over by another monitoring company is a question with a real answer, it varies by equipment and by how it was originally programmed, and it is much better established before you sign anything than afterwards.

What we touch, and what we do not

To be straightforward about the boundary, because it matters more here than on most of the jobs we do. We work on the network: the router and its configuration, the Wi-Fi the panel and its sensors depend on, splitting the bands, getting a cable or a MoCA link to where the equipment actually lives, and making sure a router swap does not orphan half the house. That is the part that breaks, and it is squarely ours.

We do not program alarm panels, we do not hold your monitoring contract, and we will not tell you your alarm is working — only your monitoring company can tell you that, because only they can see what arrives. Alarm work in California is licensed: Glendale's ordinance, to take one local example, makes it unlawful to "provide, install or maintain" an alarm system in the city "without first having obtained a license in connection therewith as required by the Department of Consumer Affairs or other state or federal agency." If the diagnosis turns out to be inside the panel, we will say so and stop.

The useful news is that the split falls heavily on our side. When somebody discovers their alarm has been quietly unmonitored since spring, the cause is a network change far more often than it is a fault in the alarm equipment.

The short version, if you do nothing else

Phone your monitoring company and ask when they last received a test signal from your system and which path it came in on. That one question costs nothing, requires no equipment, and catches the entire failure described on this page.

If you are about to change your router, provider or Wi-Fi password, keep the network name and password identical if you possibly can, and if you cannot, put the alarm at the top of the list of things to check afterwards.

And if you are going to test the alarm yourself: monitoring centre first, city police second, then trip it, then call back and ask exactly what arrived, then make sure it comes off test mode.

How we can help

If your internet has changed recently and you have a monitored alarm, cameras, solar monitoring or a house full of smart devices, this is a routine visit for us: work out what is actually still connected, get the stragglers back on, and then fix the underlying reason so the next router change does not repeat the whole exercise — usually a wired run or a stable, plainly named 2.4 GHz network rather than a reconnection that will fail again in two years.

We cover homes and small businesses across Southern California, and we are happy to say clearly when something is not ours to fix. For the alarm specifically, the last step is always the same and it is not one we can do for you: pick up the phone to whoever monitors it and make them confirm the signal arrived.