the status page no longer lives on our servers
a status page has one job: to say what works exactly when something does not. if it lives on the same server as everything else, it disappears at the worst moment too. so status.voidhosting.org is now served by a CDN (Bunny), and our servers only put two data files there.
two views
the first is our monitoring. once a minute it checks the website, the account panel, the API, the bot, the server nodes, the proxy, web hosting, domain registration and balance top-up. it counts an outage after three bad checks in a row, about three minutes: one or two failed checks also happen when the network next to the checker stumbles.
the second is a probe on another host and another network. once a minute it opens what a customer opens: the website, the account panel, the API and the sites on web hosting. its answer can only soften ours, never make it worse:
| ours | outside | on the page |
|---|---|---|
| works | anything | works |
| unstable or down | works | works |
| unstable or down | unstable | unstable |
| down | down | down |
| anything | nothing fresh | as we see it |
why. on 13 November our server lost its direct path to the hosting, while customers' sites opened all along. for a few hours the page said «unstable» although everything worked for customers. now the view from outside gives the right answer in such a case.
when our monitoring is silent
if our monitoring's file has not been updated for more than three minutes and the probe outside is fresh, the page says so: «our monitoring is not answering». under it, what is seen from outside: whether the website, the account panel, the API and the hosted sites open. the rest honestly says «no data».
if both files are old, the page says «the data is out of date» and when the last check was. it never shows a green «everything works» without fresh data.