trace a request through a configurable backend and view potential bottlenecks/issues based on the graph of the services
The organization of the nodes and the time-per-node is configured on the left of the UI.
All packages can be installed with yarn install in the root of the project (after running nvm use to select the appropriate node version). The backend and frontend workspaces need to be kicked off with yarn start.
The UI allows a creation of an arbitrary (except single entry point) directed acyclic graph of nodes which dynamically spins up several linked express servers that do an a set amount of "work" (aka sleeping) before calling down stream services and eventually returning their outputs.