“Have you tried turning it off and on again?” is a cliche because it genuinely works more often than it has any right to. But a reboot is a reset, not a repair, and if you find yourself doing it every few days, the reboot is not the solution, it is a symptom. Here is what actually happens when you power-cycle a router, and how to tell a one-off from a warning sign.
What a reboot actually does
Your router is a small computer, and like any computer it can get itself into a muddle. Switching it off and on again quietly fixes several things at once:
- It clears the memory. Routers have limited memory that fills up over time. Restarting wipes the slate and frees up whatever had clogged.
- It renews the addresses (DHCP). Devices are handed a network address on a lease. A reboot hands out fresh ones, clearing up clashes where two devices ended up fighting over the same address.
- It re-establishes the connection to your provider. The router logs back in to the network and pulls a fresh public address. A stale or dropped session that was blocking traffic comes back cleanly.
- It flushes the connection tables. Every active connection uses a slot. Busy networks can fill those slots, and a restart empties them so everything can flow again.
- It re-syncs the line. The router renegotiates its link with the exchange or fibre network, which can shake off a marginal, glitchy connection.
So why does it keep coming back?
Because a reboot treats the symptom, not the cause. If you are restarting the router regularly just to stay online, something underneath needs attention:
- Tired or overheating hardware. Ageing routers get flaky, especially in hot cupboards. If a device needs a daily reboot, it may be on its way out.
- Firmware that needs updating. Known bugs cause slow memory leaks and dropouts that only a proper update fixes.
- A marginal line. A connection that keeps re-syncing points at a line fault, not a router mood, and it will keep happening until the line is put right.
- An overloaded router. Too many devices and connections for the kit means the tables fill again within days. That is a sizing problem, not a glitch.
- Address or configuration clashes. Recurring conflicts suggest the network needs setting up properly rather than rebooting around.
- Provider-side session drops. Sometimes the connection is being reset upstream, which a pattern in the timings will reveal.
The rule of thumb. One reboot to recover from a rare blip is completely normal. Rebooting on a schedule to keep the internet alive is not a fix, it is a recurring fault wearing a disguise, and it deserves investigating.
Why it works on Wi-Fi, specifically
A lot of “the Wi-Fi was dreadful, I restarted the router and now it is fine” has nothing to do with memory or connection tables. It is about which channel you are sitting on.
Most routers scan the airwaves when they start up and pick whichever channel looks quietest at that moment. If a neighbouring business had drifted onto the same channel as you, the reboot quietly moved you off it, and the improvement is real. The catch is that the scan happened once, at the instant you rebooted, and the airwaves around you keep moving. That is why this particular fix tends to last a few days and then quietly undo itself.
If that pattern sounds familiar, the answer is not a better reboot routine. It is a proper channel plan, or access points that keep managing it for you instead of guessing once at boot. Our guide on whether it is your Wi-Fi or your internet helps work out which half of the problem you actually have.
The one time rebooting actively makes things worse
This one is worth knowing, because it is the opposite of harmless and almost nobody is told about it.
If you are on a copper based broadband line, which means ADSL, FTTC or SoGEA rather than full fibre, the network watches how stable your line is and adjusts it to suit. A line that keeps dropping and reconnecting looks unstable, so the network responds by making it more cautious: adding interleaving, which raises latency, and in some cases lowering the speed the line runs at, in order to keep it steady.
It is trying to help. The problem is that a router being power cycled several times a day looks exactly like a line that cannot hold a connection. Do it often enough and you can end up with a line that is permanently slower and laggier than it needs to be, caused by the habit rather than by any fault.
Recovery is not instant either. The line generally has to run cleanly for days, sometimes weeks, before the profile relaxes again.
On full fibre, this particular penalty does not apply. FTTP does not work the same way, so rebooting will not cost you speed. Everything else on this page still holds, including the fact that a reboot you need regularly is a fault you have not found yet.
Every reboot wipes the evidence
There is a second cost that is easy to miss. The router’s logs, error counters and connection statistics are the things an engineer would use to work out why this keeps happening, and a restart clears them.
So the common sequence, which is that it breaks, you reboot, it works, and you report it a fortnight later when you have had enough, destroys the evidence every single time. By the time anybody looks, there is nothing left to look at.
Where you can bear it, spend thirty seconds before you reboot: note the time, look at what the lights are doing, and check whether it is everything or just one device. Then reboot. That small habit is often the difference between a fault that gets diagnosed and one that runs for months.
What to do if it keeps happening
Start logging it: note the date and time each time it drops, and glance at the router lights when it does. A pattern in the timings, or a particular light behaving oddly, tells an engineer far more than “it keeps going”. When you report it, bring those details along using our fault checklist, so it gets treated as the recurring problem it is, not a fresh reboot request.
How Synergy Max helps
We monitor customers’ connections, so a line that keeps re-syncing or a router that keeps choking often shows up on our side before it becomes your daily ritual. We can update or replace tired kit, right-size an overloaded router, and chase a marginal line properly. And where being offline stops you trading, we pair the connection with failover so a blip never becomes a standstill.
Frequently asked questions
Once in a while to clear a rare blip is completely normal and nothing to worry about. Once a week is a pattern. Once a day, or any arrangement where somebody reboots it as part of opening up, is a fault that has been absorbed into the routine rather than fixed. At that point the reboot has stopped being a fix and started being a workaround with a cost attached.
We would generally say no. It hides the symptom that would otherwise tell you something is wrong, so a failing router or a marginal line can go unnoticed for months. On a copper based line it is worse than that, because a nightly disconnection can look like instability to the network and leave you on a slower, higher latency profile. If a device genuinely needs restarting every night to stay usable, that device or that line is the thing to deal with.
It can. On a standard business broadband connection the public address is usually dynamic, so reconnecting may hand you a different one. That matters if anything is reached from outside on that address, such as remote access to cameras or a VPN that other sites connect into, because those will stop working until the address is updated. If you depend on that, a static IP removes the problem.
Because the reboot moved you onto a quieter Wi-Fi channel, and a few days later the channel is no longer quiet. Routers pick a channel when they start up and then mostly leave it alone, while everyone around you is doing the same thing on their own schedule. The lasting answer is a channel plan across your access points, or kit that manages this continuously rather than once at boot.
Start with the router, because that is where most of the state that goes stale actually lives. If the fibre box, the ONT, is showing a red light for loss of signal, restarting things is not the answer at all and it needs reporting, because that indicates the signal is not reaching the building. Power cycling an ONT repeatedly achieves very little and can mask the timing of a fault that somebody needs to see.
Sometimes, briefly, if the slowness is caused by full connection tables or a router that has run out of memory. If the speed comes back for a while and then degrades again, the reboot is masking something else: an undersized router for the number of devices, a line fault, congestion at a particular time of day, or something on the network using far more than its share. Those all leave different fingerprints, which is why the timings are worth writing down.