NEVERENDERSNetwork status

Status / Method

Method

What is measured, how often, and what these numbers cannot tell you.

The probe

One HTTPS GET per domain, following up to three redirects, with a five second connect timeout and an eight second total timeout. Certificate verification is on: a domain serving a bad certificate is a failure, not a pass.

Results are cached for fifteen minutes in a file above every docroot. Without that cache this page would turn every visitor into a burst of requests against the group's own sites, which would be a strange thing for a status board to do.

The states

What these numbers are not

They are a single vantage point on shared hosting, not a distributed monitor. Response time includes this server's own network path and will be worse than what a visitor elsewhere sees, or better, and there is no way to tell which from here.

This is not an audit and it is not a service level commitment. It is the group measuring itself and publishing the result, including the unflattering rows.

No manual override

There is no admin switch to mark something green. The only way a row turns green is for the probe to succeed. A status page that can be edited by hand during an incident is not a status page, it is a press release.

Machine readable

The same data is at /status.json, with the timestamp of the measurement.