A maintenance window is a planned block of time for installing updates, patches, firmware and network changes. At a dealership that’s open every day, pick it around each department’s real hours, test first, tell the store ahead and keep a way back.
Last updated: November 3, 2026
Microsoft puts out its security updates on the second Tuesday of every month, usually around 10 in the morning Pacific time. I don’t think anybody at a dealership has ever circled that date. That evening the sales floor is still working deals, and by the time the service drive opens the next morning nobody there is thinking about Redmond. Why would they? There isn’t an hour when the whole store goes quiet. There never has been.
Those updates still have to go in, though. Firewalls and switches need their own firmware updates too, and now and then there’s a network change that hits every rooftop in a group at once. None of it is optional. Skip them and you’re carrying risk you didn’t have to carry. Push them at the wrong moment and you’ve shut the store down yourself, which is a hard thing to explain to a general manager.
I’ve been leading managed services teams for more than 25 years, and one of the teams I lead at Helion is Proactive Managed Services. They’re the people who do this planned work for the stores on our managed IT services for dealerships. So this is how I think about timing it at a store that never really closes. I wrote it so you could hold any provider to it. That includes us.

What a Maintenance Window Actually Is
A maintenance window is an agreed, recurring period when an IT team is allowed to make planned changes that might interrupt a system, like a patch that needs a restart or a firewall update that drops the connection for a few minutes. Everyone it affects knows it’s coming and roughly how long it’ll take.
The part people skip is the agreeing. It’s the important part. Plenty of shops treat the window as whenever the technician has a free hour, and the store finds out when its computers restart in the middle of the afternoon. That isn’t a window. A window is something the store signed off on and wrote down, and the people who’d notice a problem know it’s coming.
ITIL is the service management framework I’m certified in, and it calls this change enablement. The monthly workstation patches would be what ITIL names a standard change, low risk and approved once ahead of time. Swapping out a core switch is bigger, so somebody reviews it before it goes on the calendar. And then there’s the patch for a flaw attackers are already using, which can’t sit and wait for next month. Three different animals. They need three different windows, and a lot of trouble starts when someone treats one like another.
Get the type right first. The timing usually follows.
Why a Store That’s Open Every Day Has No Obvious Quiet Hour
Think about how an office building works. It empties out Friday evening and nobody’s back until Monday. A dealership doesn’t get that. Saturday is often its busiest day. Service might open before the sun’s up while the sales floor stays lit until well into the evening, and plenty of stores open on Sunday too. Every department keeps its own clock.
I’m not going to tell anyone how to run those departments. It isn’t my area. What I do need is to know when each of them is working, because the IT work has to fit around them and not the other way around. A firewall restart at 7 in the morning doesn’t look like much from an office chair. Ask a service advisor with four customers in line to drop off their cars. It’s the minute nobody can open a repair order.
Dealer groups make it harder again. Ten rooftops spread over two time zones won’t share a quiet hour, and a heavy truck branch that keeps different hours from the car store across town won’t line up either. It gets complicated fast.
So I don’t ask when the store is closed. I ask when each part of it will care the least, and who’s going to be around if something goes wrong.
The Quietest Hour Isn’t the Emptiest Hour
It’s tempting to think the best window is simply whenever the fewest people are in the building. You schedule the restarts for the middle of the night, let them run and go home. It seems efficient. Here’s what’s wrong with it.
Say a patch goes in at 3 a.m. and quietly breaks the printer driver. It happens. Nobody knows. Then the first service advisor of the day tries to print a repair order with a customer standing right there, and that’s when you find out, hours after the change, at about the worst moment you could have picked. So what problem are we actually trying to solve? It was never about avoiding people. It’s about making the change, seeing that it worked and backing it out if it didn’t, all before the store pays for it.
I’d want a different kind of window, then. The routine, low-risk updates can still install on a schedule by themselves. Fine. But the window ought to close while somebody can still print a test page and open the DMS and see that it all came back the way it should. The bigger changes go in a time when the engineer doing the work and the person at the store who’d spot a problem can both pick up the phone.
I’ll take right and secure over fast. Every time.

How a Provider Should Pick the Window
None of this is complicated. It’s just easy to skip a step when the patch looks small and everybody’s busy. This is what I believe a good process includes, whether it’s our team doing it or your own.
- Write down the hours each department and each rooftop really keeps. Memory is not a schedule, and the business office’s month-end close goes on that list as a week when nothing accounting depends on gets touched.
- Sort the change before you schedule it.
- Test on a few machines, or on one store, before everyone gets it. When a scanner driver breaks on three computers in the used car office, that’s an annoyance. On every computer in the group it’s a very long morning for a lot of people.
- Groups get staggered, one rooftop first and the rest after it’s clean.
- Does the store know? A short note that says what’s happening and who to call beats a long technical one nobody reads.
- Know the way back before you start. For a firmware update or a server change, that means a current backup and a clear undo.
- Finish while people are there to confirm it worked, and close it out with a note.
The way back is the step people cut first. I’d keep it over anything else on that list. Our post on backup and disaster recovery for dealerships covers the backup side in more detail.
Network changes deserve their own care. A switch or firewall update can take every device at a rooftop offline for a few minutes, including phones and card terminals. That’s work for a reviewed window with an engineer who knows the store’s layout, which is a big part of what our dealership network management team does.
| Type of change | Usual impact on the store | When it fits | Who signs off |
|---|---|---|---|
| Monthly workstation patches | A restart per computer, a few minutes each | A recurring window after each department’s busiest hours, tested on a small group first | Pre-approved once, as a standard change |
| Server updates and restarts | Shared files, scanning or local apps unavailable for a while | A scheduled window with someone able to check the result before the store depends on it | The store’s IT contact or general manager |
| Firewall, switch or Wi-Fi firmware | The whole rooftop offline briefly, phones and card terminals included | A reviewed window, one rooftop at a time, with a tested way back | General manager or dealer principal |
| Changes near accounting systems | Risk to month-end close and payroll | Outside the month-end close | Controller or office manager |
| Emergency security patch | Varies; waiting is the bigger risk | As soon as it’s tested enough to trust, with the store told why | A named contact who can approve on short notice |
Every change needs an owner and an undo.
Two Ways It Goes Wrong
The first is the one everybody pictures. Someone pushes an update in the middle of a busy Saturday because it was sitting in a queue, a restart hits the F&I office while a deal is being signed, and the customer waits while a computer reboots. Nobody meant any harm. The window just wasn’t agreed, so it was whenever.
The second is quieter and, I think, more common. The updates get put off. There’s always a reason: it’s the end of the month, a big sale is on, the last patch caused a problem, nobody wants to be the one who broke the store. Weeks turn into months. What happens if we do nothing? The store keeps running on software with known holes in it, and the eventual catch-up is bigger and riskier than all the small updates would have been.
That second pattern is the one I’d worry about. The National Institute of Standards and Technology’s Guide to Enterprise Patch Management Planning (SP 800-40 Rev. 4, 2022) frames patching as preventive maintenance for technology, the same way a service department treats oil changes. You don’t skip them because the car’s still running. The U.S. Cybersecurity and Infrastructure Security Agency keeps a Known Exploited Vulnerabilities catalog, a list of flaws attackers are already using, and says organizations should use it to decide what gets fixed first. A flaw on that list is an emergency change, not something that waits for next month.
The cost of getting it wrong is real. CDK Global’s 2023 State of Cybersecurity in the Dealership study put average downtime after a dealership cyberattack at 3.4 weeks. A few minutes of planned restart time is a small price next to that.
We’ve also written about the older side of this in software security patching as a hole in dealership cybersecurity, and about machines that are too old to patch at all in our guide to the dealership PC refresh cycle.
Late and rushed are both risks. Planned is the fix.

Who at the Store Signs Off
We can’t pick the window on our own, and I don’t think a provider should try. The store knows things we don’t. There’s a tent sale this weekend. The controller is closing the month on Thursday. Service is starting an hour earlier for the summer. None of that shows up in a patch schedule unless somebody at the store tells us.
It’s fairly simple. The general manager or the dealer principal approves the overall plan once. Each department manager tells us which hours are off limits for their area. And there’s one named person who can say yes to an emergency change on short notice, which in a dealer group usually means someone at the group level, so one call covers every rooftop.
Then write it down. One page that says when the windows are and who to call in an emergency will do more than a lot of good intentions, and it keeps working after a manager moves on. It’s the same reason we push for IT documentation that lives in systems rather than in someone’s head.
Hold the provider to it. When a change is made, the store should hear what was done and whether it worked. If something did break, people should be able to reach us through our dealership IT help desk, and the ticket should show it came from a planned change so the fix starts in the right place.
The plan is the store’s. The work is ours.

What Dealers Ask About Update Timing
Can’t updates just install on their own after the store closes?
How often should a dealership actually be patching?
The last patch broke a printer. Why not just stop patching until things settle down?
We’re one rooftop. Do we need a formal maintenance window?
Is Sunday a good day for updates?
Plan the Change, Then Make It
Updates aren’t optional, and neither is the store’s day. The work is finding the time when a change costs the store least and can still be checked, then doing it the same careful way every month. That’s not glamorous. It’s doing things correctly and securely, and the people on our Proactive Managed Services and engineering teams are the ones who make it hold up, store after store, across the 2,000-plus dealerships we support.
If you’re not sure when your own stores get updated, or whether something is quietly months behind, our complimentary IT and cybersecurity assessment will show you where things stand. Start there, then agree on a window you can keep.
A good window is one you planned ahead and can undo.

