@mtth/stl-bootstrap

Opinionated wiring for spinning up telemetry in Node.js services. Create an AppTelemetry once and get logs, metrics, traces, and context propagation that obey the OpenTelemetry environment variables out of the box.

Quickstart

pnpm add @mtth/stl-bootstrap

import {AppTelemetry, enableContextPropagation} from '@mtth/stl-bootstrap';
import {instrumentsFor} from '@mtth/stl-telemetry';

enableContextPropagation(); // idempotent, safe to call at startup

const telemetry = AppTelemetry.create({name: 'orders-api', version: '1.0.0'});

telemetry.logger.info('service bootstrap complete');
const [metrics] = telemetry.metrics(
  instrumentsFor({requests: {name: 'http.server.requests', kind: 'counter', unit: '{requests}', labels: {result: 'result'}}})
);

process.on('SIGTERM', async () => {
  telemetry.logger.info('Shutting down...');
  await telemetry.shutdown();
});

Setting OTEL_EXPORTER_OTLP_ENDPOINT (or the metrics/traces-specific environment variables) automatically turns on the corresponding exporters.

Highlights

• AppTelemetry.create() provisions logger providers, metric readers, and span processors with a single call, including graceful shutdown hooks.
• Built-in resource detection reads the standard OTel environment variables so metrics and traces share consistent service metadata.
• Context propagation (AsyncLocalStorageContextManager) is enabled for you, making spans flow automatically across async boundaries.
• Extra helpers such as PrometheusMetricsReader and prometheusMetricsSerializer make it trivial to expose on-node metrics.