Crate o_sfu_telemetry 

telemetry setup, metric catalog and diagnostics schema

o-sfu-telemetry keeps runtime observability contracts in one crate the server initializes tracing from TelemetryConfig, records process-local metrics through typed helpers and renders diagnostics or Prometheus output from the metric catalog

use o_sfu_telemetry::{
    metrics::{HttpRoute, RoomGaugeValues, RuntimeMetrics},
    prometheus::render_prometheus,
};

let metrics = RuntimeMetrics::default();
let request = metrics.track_http_request(HttpRoute::Noop);
let body = render_prometheus(&metrics, RoomGaugeValues::default());
drop(request);
assert!(body.contains("osfu_http_noop_requests_total"));

process-global tracing setup lives behind init_tracing code that only needs counters, diagnostics or schema names can use those modules without installing a tracing subscriber

Modules

config 🔒

diagnostics

graph

Grafana node graph formatting for diagnostics snapshots.

metrics

prometheus

Prometheus text rendering for runtime metrics.

schema

telemetry event and field names

setup 🔒

Structs

TelemetryConfig

TelemetryHandle

TelemetryResource

TraceExportConfig

Enums

TelemetryLogFormat

Constants

DEFAULT_MEDIA_QUALITY_INTERVAL

DEFAULT_TELEMETRY_DEPLOYMENT_ENVIRONMENT

DEFAULT_TELEMETRY_SERVICE_NAME

Functions

activated_span

http_request_span

init_tracing

Errors

ws_handshake_span

ws_upgrade_span