@mtth/stl-resilience

Deadlines, retries, and health instrumentation that wrap any async operation with observability-friendly guardrails.

Quickstart

pnpm add @mtth/stl-resilience

import {resilient} from '@mtth/stl-resilience';
import {RecordingTelemetry} from '@mtth/stl-telemetry';

const telemetry = RecordingTelemetry.forTesting({name: 'orders-api'});

const fetchOrder = resilient('fetch order', {retryCount: 2}, async ({span, deadline, logger}) => {
  logger.debug('Calling upstream...');
  const res = await fetch('https://example.internal/orders', {signal: deadline.signal});
  if (!res.ok) {
    throw new Error('Upstream failed');
  }
  span.addEvent('upstream complete');
  return res.json();
});

const order = await fetchOrder.run({telemetry});

Highlights

• Wrap any promise-returning function with deadlines, abort signals, retries, and OpenTelemetry spans via the resilient() helper.
• Listen for retry/lateAttempt* events to feed alerting systems or emit extra metrics when work keeps failing.
• Deadline and race helpers compose with AbortController, letting you enforce cascading timeouts across API layers.
• Built-in settings + telemetry integration expose knobs (log levels, backoff windows, retry codes) without duplicating boilerplate in every service.