A deployment is a version of an application running in an particular environment, e.g stage, production, canary. Deployments can contain multiple traced operations on multiple worker nodes.


An operation is any function, event or action that is traced by Graphsignal. Operations represent interfaces between components or services, e.g. request handlers, API endpoints, or other significant functions.


A trace represents a graph of operations or events in context of user request, AI agent run or model inference, with details such as timeline, errors, data, and resource utilization.


Session groups multiple traces together to represent a single run, thread, conversation or user interactions.


Scores are numerical values and metadata assigned to events or objects identified by tags, e.g. generations, runs, sessions, or users to evaluate their behavior, importance, performance or impact. For example, scores are useful to analyze and act on negative user interactions.


Metrics show different aspects of execution and data in time. Graphsignal provides performance, system, data, and cost metrics.


Profiling allows recording code-level and kernel-level execution statistics. Graphsignal can use various profilers to enrich traces.


Graphsignal tracer is a module that is integrated into applications and scripts. It exposes tracing API, computes performance and data statistics, and uploads data to Graphsignal servers.