In most cases, a single Host Agent is installed on the StackState server to provide Agent-less integration with APIs from multiple sources. Data is gathered and received by one or more Agents and delivered at the Receiver API. From there, all data is put on Kafka. The data is processed by microservices and ends up as Topology in our versioned graph database, called StackGraph. Traces, and some telemetry data, are temporarily stored in Elastic Search.