How we check

What this page measures, how, and what it can't see.

Checking…

What each state means

Operational
Working normally.
Degraded performance
Working, but slower than normal.
Partial outage
Part of it isn't working, or it isn't working for some players.
Major outage
Not working.
Under maintenance
Switched off on purpose while we work on it.
No data
Not checked in this period.
Not in use
Skylith doesn't use it yet.

What we check

Our server asks for every service’s pages at skylith.io, the way your browser does, so each request passes through Cloudflare first. That’s how the page tells a Cloudflare security rule blocking a page apart from our own servers failing. It also checks each backend system directly, like the database, authentication, media storage, bot protection and email.

Cloudflare’s bot protection can show our checker a browser check, because the checker is a program. Then the same page is checked at Vercel too, past Cloudflare. If it can’t be, the service shows No data and says why.

A page that takes more than 2 seconds is slow, and so is the database or authentication after 1.5 seconds. No answer within 4 seconds counts as down. A failed check is asked again before it counts, so a dropped connection isn’t an outage.

When and from where

About once a minute while this page is open, every five minutes while people are using Skylith, and once a night. The checks run from Washington, D.C., in the US, so a Cloudflare rule that only applies to other countries won’t show here.

A check speaks for up to 5 minutes either side of it, so checks up to 10 minutes apart cover all the time between them. Time no check covered, like hours with nobody on Skylith, shows as No data.

How uptime is worked out

Uptime is the share of the checked time a service was working. A major outage counts fully against it, a partial outage 30%, and planned maintenance not at all. Time with no check doesn’t count either way.

Each bar is one day, midnight to midnight UTC. It shows the worst state at least two checks found that day, or an incident. Point at a bar, or move along it with the arrow keys, to see how long that day spent in each state.

What this page can’t see

This page runs on the same servers it checks. If Cloudflare or our hosting is completely down, this page may not load either. When it loads but can’t get the status, it says who refused the request.

A problem on your own device or network won’t show here. If Skylith isn’t working for you while everything here is operational, tell us at support@skylith.io.

Incidents

When 2 checks in a row find a service down or partly down, this page opens an incident by itself, and a bar across Skylith says so. When 2 checks in a row find it working again, the incident closes. Our team posts the problems checks can’t see, like a bug in the game.

Detected
Our checks found it. It closes by itself once they find it working again.
Investigating
We know something is wrong and are finding out why.
Identified
We know the cause and are fixing it.
Monitoring
A fix is in. We're watching to make sure it holds.
Resolved
It's fixed.

Planned work goes through Detected, Scheduled, In progress, Verifying and Completed.