Freeze Budget Explainer

what to do

• Move long timers to a Service Worker. SW setTimeout is governed by the same budget, but Service Workers are designed for periodic wake (push, periodic sync).
• Persist state proactively. Treat any backgrounded event as “might be the last thing you ever run”. Use visibilitychange + pagehide as a save point.
• Use Background Sync for “run when next online” tasks instead of polling.
• Use fetch(url, { keepalive: true }) for departure beacons — they survive freeze. Pair with the new keepalive retry in v139.

snippet

// Save before the tab might freeze.
document.addEventListener('visibilitychange', () => {
  if (document.visibilityState === 'hidden') {
    fetch('/save', {
      method: 'POST',
      body: JSON.stringify(draftState()),
      keepalive: true,           // outlives the freeze
      // Chrome 139 will retry this beacon if the network drops:
      retryOptions: { maxAttempts: 3, initialDelay: 1000 },
    });
  }
});

// Replace setInterval polling with periodic sync.
const reg = await navigator.serviceWorker.ready;
await reg.periodicSync.register('refresh-feed', {
  minInterval: 24 * 60 * 60 * 1000,  // 24h
});

see also

• Feature page
• Freeze Timeline
• Lifecycle Recorder
• ChromeStatus entry
• Page Lifecycle spec