Define canonical metrics once. Ingest events from any source. Run versioned manifests over real-time streams or post-mortem datasets.
The framework is built around stable data contracts, a canonical dictionary, and reproducible manifests rather than hardcoded analysis scripts.
Metrics are defined with stable names, units, types, and semantic meaning. The dictionary is
exposed through api.pitgun.io and used by runtime contracts and processing
manifests.
Analysis and pipeline behavior is described in versioned manifests: formulas, derived metrics, filters, aggregation windows, and output contracts stay reviewable and reproducible.
Versioned JSON Schemas define the stable boundary between producers, processors, and insight services: event envelopes, metric dictionaries, pipeline manifests, bundle manifests, and analysis contracts.
Pitgun the game generates real telemetry, sessions, player behavior, and operational metrics. It exists as a demanding workload to prove the framework against live traffic.
Games, simulators, devices, or replay tools emit runtime events.
Events cross the boundary through
pitgun-envelope-v1.
Channels map back to canonical names, units, dimensions, and expected types.
Pipelines apply formulas, filters, derived metrics, and aggregation rules.
Outputs feed reports, dashboards, datasets, and downstream APIs.
Pitgun Game is a live telemetry generator: browser clients produce event envelopes, the gateway validates and stores them, and the backend turns sessions into performance data. The game is the demanding workload; the framework is the reusable architecture underneath it.