Every hosting provider on earth claims 99.9 percent uptime or better. A lot of them claim more. If you read the status pages carefully, the uptime numbers they publish do not always match the downtime you actually experienced.
Why The Numbers Do Not Match
Two reasons. The first is scope. A provider can exclude planned maintenance, partial outages, regional issues, and problems affecting only some customers, and still report a clean month. The second is the definition of down. Some providers count the service as up if the login page loads, even when mail is not flowing. Some only count downtime when more than half of their customers are affected.
Those are not lies, technically. They are choices about what to measure. If you did not choose those measurements, the number does not mean what you think it does.
What We Do Differently
Our uptime is tracked in public. The same monitoring that watches our stack is the monitoring that feeds our status page. If a customer saw an outage, the status page saw it too. We do not redact partial outages because they only affected a subset of customers.
If you ever want the raw data, we will send it. Per service, per month, with the outage windows annotated. You can build your own uptime calculation from that, using whatever definition of down matches your needs.
What Uptime Does Not Cover
Availability is only one dimension of reliability. A service can be up and still be slow, broken in quiet ways, or failing for a specific customer because of something local to their setup. We watch for those too, and our support team treats slowness and partial failures with the same urgency as a total outage.
The number to ask about is not 99.9. It is whether, when things do go sideways, you will know about it quickly and hear from us directly. That matters more than any glossy dashboard.
