All systems
operational.
The control plane, the ingestion pipeline and the runtime API are serving normally. This page is where the current state is stated plainly, without a dashboard to interpret.
During an incident
What is affected, what still works and what we are doing is posted here and updated until it is closed. Workspaces with an incident touching their own data are told directly rather than left to find it.
What degraded looks like
Ingestion falling behind delays what the graph knows, it does not lose it. A runtime a workspace cannot reach is reported as its own state in the product, and policy fails closed rather than waving actions through.
After it is over
Anything that affected decisions or approvals is written up, with what changed so it does not recur. The audit trail of what ran during the window stays intact either way.