Everything we monitor, in public.
Including the incidents that made us look bad. What you will not find here is a wall of reassuring bars we drew ourselves: until the monitoring is wired up and has run, this page publishes the inventory and nothing else.
No availability data published yet
Monitoring is not yet feeding this page. We would rather show you an empty table than a plausible one.
Why this page is empty
An uptime figure is worth exactly as much as the measurement behind it. We have not started recording one, so there is nothing here to show — no ninety-day bars, no mean availability, no incident history. The alternative was to generate numbers that looked right, which is what this page used to do and what we are not willing to publish next to an SLA that pays out in credits.
When probing starts, the first recorded day becomes the beginning of the history and every service below fills in on its own. Until then, the 99.9% in the SLA is a commitment you can hold us to, not a result we are claiming.
Network
Compute
Platform
Incident history
Every incident, with what actually went wrong and what we changed so it does not recur. Post-incident notes are written by the engineer who handled it, not by marketing.
Nothing published yet
No incident report has been published. That means none has been written up here — it is not a claim that nothing has ever gone wrong, and you should read it as an empty log rather than a clean one. Three invented post-mortems used to sit in this space, complete with timestamps and named upstreams, and deleting them cost this page its most convincing section. That was the right trade.
Get notified
Incident updates by email, Telegram or webhook. Configured in the control panel, on by default for services you own.