Sever the link between a backend service and its database in our new Playground, and a change event lands in the graph. A few seconds later the root cause comes back: the exact link you cut, and why the service can no longer reach its data!
That's the same move Annie makes in production, and now you can drive it yourself in the browser. No signup, nothing to install.
What you're actually doing
The Playground is a split screen. On the left, the Anyshift panel where Annie reports. On the right, an interactive map of a small piece of infrastructure: services, a database, the links between them.
You break something. Sever the link between the backend and its database, or kill the database node outright. Either action records a change, the same way a real deploy or a console edit would. Annie picks up that change, investigates it against the topology, and streams back the root cause: what changed, and what it took down.
It resets on reload, and right now there's a single scenario to break. That's the honest scope: a sandbox to feel the loop, not a model of your own estate.
Why a graph and not a log
Most root-cause tools start from telemetry: logs, metrics, traces, an alert that already fired. They tell you something is on fire and leave you to work backwards to the change that lit it.
The Playground starts from the other end. The graph already knows the backend reaches the database through that one link. So when the link is cut, there's nothing to work backwards from: the change is the lead, and the topology says exactly what sits downstream of it. You feel the difference in the few seconds it takes Annie to answer, without wiring up a single integration.
Go break something, and watch the change become the answer!
