Fleet uptime reports built on three honest states, not two
CloudyFy TV reports device availability as up, down, or unknown, so a screen that has never connected is labeled "never online" instead of a misleading 0% uptime figure. Every report exports to CSV.
What do CloudyFy TV's uptime reports show?
Uptime reports answer one question honestly for every device on your account: was it actually reachable during the period you're looking at. Availability is calculated from three states rather than a simple up-or-down toggle: up (the device was reachable), down (a real offline interval was recorded), and unknown (no evidence exists yet, for example a screen that was added but never switched on). Time spent in the unknown state is never folded into the down count, so a screen you haven't commissioned yet doesn't drag down your fleet's real uptime number.
This distinction exists because a rollout, not a failure, is the most common reason a screen shows no data: a device ordered for a new location, added to the account so it's ready to configure, but not yet unboxed on site. Counting that as downtime would make every fleet look worse the moment new hardware is provisioned, which is exactly backwards for a report meant to flag real problems.
- Per-device and fleet-wide uptime, over 24 hours, 7, 30, or 90 days
- A "never online" label for devices with zero connection history, never a fake 0%
- Offline incidents, individually and collapsed into groups when outages repeat close together
- A time-weighted fleet average, not a simple average of percentages that treats every device's window as equal length
- CSV export on every report
How does uptime reporting work?
Pick your window
Choose 24 hours, 7, 30, or 90 days; 7 days is the default. Ranges longer than a year aren't accepted, since evidence that old isn't meaningful for a live fleet decision.Read the fleet summary
See a time-weighted average uptime percentage across every device, plus a separate average restricted to devices that have connected at least once, so a backlog of never-commissioned screens doesn't distort the headline number.Drill into one device
Open a device's own timeline of up, down, and unknown segments, plus its individual incidents.Review incidents
Outages within 15 minutes of each other collapse into one incident group with an occurrence count, so a flapping connection reads as one event, not fifty.Export to CSV
Pull the filtered report for a stakeholder update or an SLA conversation.
Why does honest uptime reporting matter?
No false alarms from screens you haven't set up yet
A device added last week that hasn't been switched on shows as "never online," not as a screen that's been down 100% of a 30-day window.
Outage duration is real, not inflated
A device offline since before your reporting window began is clipped to the window you asked about, so one old outage can't dwarf a 30-day report.
Flapping connections don't flood the incident list
Repeated short outages on the same device collapse into one group, so you see a pattern instead of noise.
You can defend the number
A time-weighted fleet average, with its denominator stated, holds up under scrutiny in a way an unweighted average of mismatched windows doesn't.
Honest reporting vs. a simple up/down count
A two-state uptime model has no way to represent "this screen hasn't been switched on yet" except as downtime, which quietly punishes a fleet in the middle of a rollout. It also has no way to tell a five-minute blip from a five-day outage without the raw timestamps, so incident lists become noise. CloudyFy TV's three-state model and flap-collapsed incidents exist because a fleet manager needs a number to act on, not a number to explain away.
The proof: how the fleet average is calculated
The fleet-wide average is time-weighted: the sum of uptime across every device divided by the sum of tracked time, the same quantity each device's own row reports, just pooled. It is shown alongside a second figure restricted to devices that have connected at least once, because pooling never-commissioned screens into one average conflates a rollout backlog with a reliability problem. Neither figure is shown without naming its denominator.
FAQs
Get an uptime number you can defend
Start the free trial and pull your first fleet report this week.