Cache Size Limitations
You'll learn how browser storage quotas, precache size, and cache eviction policies constrain your service worker caching strategy.
The Storage Reality
Unlike server-side storage where you control capacity, browser caches live under strict quotas. Each browser allocates a limited amount of disk space for your origin's data (Cache API, IndexedDB, LocalStorage combined), typically a percentage of available disk space (often 10-50% of free space, with per-origin limits around 20-60% of that pool).
Think of it like a hotel with limited rooms: you can't keep every guest forever, and when space runs tight, some have to check out.
Precaching and Initial Install Size
When your service worker precaches assets during installation, those files count against your quota immediately. A bloated precache manifest (loading hundreds of large images or videos) can:
- Fail the install if it exceeds quota mid-download
- Make first-time installation slow and bandwidth-heavy
- Leave less room for runtime caching
Best practice: Precache only critical shell assets (HTML, core CSS/JS, essential icons). Let runtime caching handle everything else on-demand.
Runtime Cache Eviction Policies
When quota fills up, browsers use LRU (Least Recently Used) eviction:
- Least-accessed caches disappear first
- Individual cache entries within a cache may be removed
- No warning to your service worker—data just vanishes
Your defense: Implement cache size limits and expiration policies (as covered in earlier lessons). Manually delete old entries before the browser does it unpredictably. Prioritize which caches are "best-effort" vs. critical.
Checking Available Quota
Use the StorageManager API (if needed in future builds):
// Concept illustration only - you'd check quota availability
// to decide whether aggressive caching is safe
Key Takeaway: Browser storage is finite and shared; design cache strategies that stay within quotas by limiting precache size, setting expiration policies, and accepting that runtime caches are disposable.