Ask a church tech director about their worst Sunday and the story almost always has a prologue nobody told them. The booth computer had been restarting itself since Tuesday. The overflow TV went dark a week ago and the greeters just stopped pointing people to it. The backup camera stopped answering and the volunteer who noticed worked around it, because there was a service to run.
Then it’s 9:55, the stream won’t start, and the director hears about all of it at once.
This is not a people problem. It’s a reporting problem, and it has a shape worth understanding before you try to fix it.
Why the problems stay quiet
A volunteer who finds something broken at 9:50 on a Sunday is not going to file a report. They have a service to run. They find a workaround (switch to the other camera, skip the overflow feed, restart the computer and hope) and the service happens. By noon the problem feels solved, because the thing that mattered, the service, went fine.
The next volunteer inherits the workaround without knowing it’s a workaround. “We don’t use that camera” becomes a fact about the booth. Three weeks later, the director asks why the stream has had one wide shot all month, and nobody can say when it started.
Nobody in that story did anything wrong. There was simply no moment in the morning when reporting was the natural next step, and no way to report that didn’t feel like complaining.
The quiet week is the dangerous one
The booth is at its most fragile when nobody is in it. Windows updates itself on a Wednesday night and reboots into a login screen. A disk fills up with recordings. A network switch loses power when someone unplugs it to charge a phone. A camera’s PoE injector dies on a Thursday.
None of those are Sunday problems when they happen. They become Sunday problems at 9:55, because nothing was watching between Sunday and Sunday. The fix is not a more diligent team. It’s something that looks at the booth on Wednesday night so a person doesn’t have to.
The three signals that catch almost everything
Churches that stop being surprised on Sunday morning are watching three things, and they’re letting a system watch instead of a person.
The booth computer. Whether it’s online, and how it’s doing: CPU load and temperature, memory, every disk and how full it is, the GPU and its driver, the displays, the audio endpoints with their volume and mute state, the USB peripherals. A computer that’s been restarting since Tuesday shows up as a computer that went offline on Tuesday. With Pulse, a telemetry beat lands every fifteen seconds, and a machine earns its offline badge after ninety seconds without one, so the Wednesday-night reboot is caught on Wednesday night.
The devices on the network. The cameras, the console, the switcher. A device you’ve flagged for monitoring alarms when it stops answering, and only when every booth computer that probes it agrees, so one booth’s bad network path never pages the whole team for nothing. The dead PoE injector on Thursday becomes an alert on Thursday, not a shrug on Sunday.
The stream, and the sound going into it. Whether it’s still running, the bitrate and dropped frames while it runs, and the classic pre-service mistake nobody can see from the booth: the computer’s default output muted, or under ten percent, or the microphone muted. That one is raised as a warning before service and cleared when it’s put right.
Alerts that go to a person, once
Watching is only half of it. The other half is that someone hears, and hears once. An alert that fires forty times for one flapping machine trains everyone to ignore it, which is worse than no alert at all.
The useful shape: an outage alerts exactly once per incident. Each kind of alert goes to the people who switched it on for themselves, by email and the booth notification, and by text on Pro. A critical computer alert lands in the inbox for everyone; a quieter one, like a booth going offline overnight, notifies only the people who asked for it. Each card can be acknowledged, resolved, or muted, so the inbox stays honest.
The director reads the alert at the breakfast table on Sunday, or on Wednesday night, and the first text from a member at home never gets sent.
What the team still has to do
Monitoring doesn’t replace the people in the booth. It replaces the reporting they were never going to do. The team still runs the Sunday morning checklist, still walks the room, still watches the stream. What changes is that the checklist gets shorter, because the things that used to require a human eye (is the computer up, is the audio muted, is the camera answering) are already known before the first volunteer arrives.
And when something does break, the person helping can see it: support that opens with your system on screen, instead of a description from memory while the countdown runs.
Next: see what a tech director knows at 8:40 am before leaving the house, from System Health to the one-tap restart, on the Sunday at a glance page.
Related questions
- Why don't church volunteers report tech problems?
- Usually because reporting is work and working around it is easier. A volunteer who finds the backup camera dead at 9:50 has a service to run; by noon the problem feels solved because the service happened. Nobody is hiding anything. There is simply no moment in the morning when reporting is the natural next step.
- How can a church tech director find out something is broken without being told?
- By letting the system report instead of the people. Monitoring that watches the booth computer, the devices on the network, and the stream whether or not anyone has a page open will raise an alert the moment a computer goes offline or a device stops answering, and send it to the people who switched that alert on.
- What should a church monitor in the booth?
- The booth computer (CPU, memory, disk, temperature, and whether it is online), the devices that matter on Sunday (cameras, consoles, switchers), the stream (bitrate, dropped frames, whether it is still running), and the classic pre-service mistake: the booth computer's default output muted before anyone notices.
A Church Tech Director's Sunday Morning, 8:40 to 10:00
NextStop Titling Church Livestreams 'Live Stream' and a Date: A Naming System That Makes Your Catalog Findable
Related Articles
Build the Church Tech Schedule in 20 Minutes: Positions First, Then People
The schedule that takes all week is the one that starts with people. Start with positions, let availability narrow the list, and invite one person per slot. A month of Sundays in twenty minutes.
What Church Leadership Should See on Monday: Attendance Trends That Mean Something
A single Sunday number tells leadership almost nothing. Here's what to put in front of the elders on Monday instead: in-person and online together, total reach per service, the trend by month and year, and the one list that turns a number into a phone call.
One Sunday, Four Audiences: The Lobby, the Room, the Feed, and the Living Room
Every Sunday a church talks to four audiences on four kinds of screen, usually with four tools and four people. Here's what each audience actually needs, and how one platform covers all four without a media team.