Uptime and availability
Every device, gateway and monitor on your account has an availability history. You can see it as a percentage over a period, as a day-by-day strip, or as a list of the outages behind it.
Up, down — and unknown
Most uptime tools show two states. Synacl shows three, and the third one matters.
- Up — we heard from it, or a probe reached it.
- Down — it stopped reporting past its expected interval, or a probe failed.
- Unknown — we have no evidence either way for that stretch of time.
Unknown time happens for real reasons: the period predates the device, it predates the history your plan retains, or the platform itself had a gap. Unknown time is never counted as up. It is left out of the calculation entirely rather than quietly assumed to be fine.
That is why every availability figure is shown with a coverage percentage next to it. Coverage says how much of the period the number is actually based on:
99.8 % available · based on 62 % of the period
Those two numbers together are honest. The first one on its own would not be — a "99.8 %" built on two days of evidence out of thirty is not something you would want to rely on, and it is certainly not something you would want to put in front of a customer.
When there is no evidence at all, you will see —, not 0 % and not 100 %. Neither of those would be true.
How far back you can look
Availability is stored as one small record per source per day, so history survives long after the underlying events have expired.
| Plan | Uptime history |
|---|---|
| Free | 7 days |
| Starter | 30 days |
| Pro | 90 days |
| Business / Enterprise | 365 days |
Ask for a longer period than your plan allows and you get the longest window it does allow, with a note saying so — not an error.
If you upgrade, your older history is still there. Records are kept regardless of plan; your plan controls how far back you can read, not how much is stored. Downgrading hides history rather than deleting it.
Where it comes from
Uptime is reconstructed from the same online and offline events you see in your event feed
(device/online, device/offline, gateway/online, gateway/offline). Nothing separate is
measured, so what you see on the uptime strip and what you see in the feed always agree.
A device is considered offline when it goes quiet for longer than its expected reporting interval allows — roughly three missed reports, with a floor of 90 seconds and a ceiling of one hour. See Devices for how that interval is set per protocol.
What to do with it
- Put an uptime bar strip on a dashboard for the handful of things you care about most.
- Use an SLA widget when you have committed to a number and want to watch the error budget.
- Check the incident log after a bad day — it shows each outage with its start, end and duration, stitched correctly across midnight so one outage is one row.
- Add a monitor for anything that isn't a Synacl device: a website, an API, a server, a shop's point-of-sale box.
Set up a rule on device/offline if you want to be told rather than having to look.