All systems operational

Every component on this page is probed once a minute from our own infrastructure, and each probe result is written to the day it lands in. There is no manual switch: nobody marks a component green by hand, and an outage appears here whether or not anyone has noticed it yet.

The components are probed independently because they fail independently. Website and Dashboard cover the public site and the Client Area sign-in. API covers the public REST endpoints. Downloads checks that the build-delivery endpoints are answering. It confirms the service is up, not that a specific customer's build has been produced, since a probe has no licence to authenticate with. Plugin Connectivity is the endpoint your servers dial out to, and Detection is the live analysis of the gameplay they report. The last two are the ones that matter while a match is running.

A component is degraded when it answers but not within the budget we expect, and in outage when it does not answer at all. A day with no samples is drawn blank rather than green and is excluded from the average. A gap in our monitoring is not evidence of uptime, and this page will not round it into one.

If Detection or Plugin Connectivity is down

Your server keeps running. The plugin dials out to us rather than the other way round, so a cloud-side outage cannot hang or crash your server, and it does not stop players from joining or playing. What pauses is the cloud-side analysis: checks that depend on our models stop producing verdicts until the link is back, and the plugin reconnects on its own once it is. You do not need to restart anything.

If your server cannot connect while this page reports everything operational, the problem is almost always egress filtering on your host rather than an incident here. Run /ecstacy doctor in game, which works even with the cloud link down, and see troubleshooting for the specific console messages and what each one means.

Incidents

Anything affecting customers gets an entry under Recent Incidents with the components it touched and a timeline of updates, including after it is resolved. Incidents are not deleted once they close. The history stays on this page so the record of what broke, and for how long, remains checkable.

How the uptime figure is calculated

Each bar is one day: successful probes over total probes, for that one component. The headline percentage averages the days that have samples. Days without any are left out of both the chart and the average instead of being assumed good, which is why the number can shift slightly as a day fills in.

Probes run from our infrastructure, so what this page measures is whether the service answered us. It cannot see a network path that is broken only between your host and us. Everything green here while your server still cannot connect is itself the answer.

Checking it from your side

Three commands, in this order:

• /ecstacy cloud gives you connection state, lobby host, ping and last heartbeat. A recent heartbeat means the link is alive right now.
• /ecstacy doctor is the full diagnosis. It runs with the cloud link down, which is exactly when you need it. Support asks for this output first, so paste it as-is rather than summarising.
• /ecstacy suspicious lists players scoring near the detection bar. Populated means analysis is flowing. Empty on a busy server points back at the link.

Being told before you notice

This page is public and needs no account, so an external uptime monitor can watch it and your staff docs can link it. For programmatic checks the HTTP API exposes activeInstances for your licence on /api/v1/info. That is the more useful alarm: it drops when one of your servers loses its link, which a global status page cannot tell you.

Incidents are posted on Discord as they open and update. Fastest way to see if something is already known.