Recent outages across everything we monitor
Every incident our checks confirmed in the last 30 days, newest first. Times are UTC. Subscribe by RSS to get these as they close.
Nothing recorded in this window
No monitored service failed two consecutive checks in the last 30 days.
What counts as an outage here
Every entry in this log was confirmed by two consecutive failed checks from our own servers. An outage starts at the first of those failures and ends at the first successful check afterwards, so the durations are measured rather than estimated or reported.
Why these durations differ from official figures
Most published availability figures exclude announced maintenance, third-party faults and anything a provider classes outside its control. Ours exclude nothing. If the service did not answer, it is in this log. That makes our numbers consistently less flattering and consistently comparable between services, which is the trade we chose. You can see what any uptime percentage actually permits with the uptime calculator.
What the log cannot tell you
- Cause. The recorded cause is what our probe saw at the moment it failed. A timeout is a symptom, not an explanation, and we never guess at the rest.
- Scope. We measure from one network. An outage listed here may have been global or may have affected only the route between us and them.
- Impact. Thirty minutes at three in the morning and thirty minutes at peak are the same number here and are not the same event.
Why an archive like this gets more useful over time
A single month of measurement says very little. Several years of it, recorded the same way for every service, is the only honest basis for comparing one provider with another - and unlike a status page, it cannot be edited after the fact. Each service’s own record is on its history page, reachable from any name in the table above.