Skip to Content
APIConfigurationInterface: DeckTileCacheConfig

Interface: DeckTileCacheConfig

In-memory cache bounds for the deck.gl MVTLayer overlays (labels, POI icons, store logos, elevated lines). deck.gl keeps every parsed tile — and its GPU buffers — in memory unless a bound is set, so on a long-running kiosk that pans/zooms/changes floors for hours the working set grows without limit. These bounds cap what stays resident in RAM/VRAM; the raw tile bytes remain in the Service Worker cache on disk, so an evicted tile is re-parsed from disk rather than re-downloaded.

All fields are optional and forwarded verbatim to every MVTLayer. When omitted, the SDK applies conservative defaults (see DEFAULT_DECK_TILE_CACHE) — the overlays are always bounded, never unlimited.

Properties

maxCacheByteSize?

optional maxCacheByteSize?: number;

Max total byte size of parsed tiles kept per overlay layer (deck.gl maxCacheByteSize). Only enforced when deck.gl can determine a tile’s byte size; use together with maxCacheSize.


maxCacheSize?

optional maxCacheSize?: number;

Max number of tiles kept parsed in memory per overlay layer (deck.gl maxCacheSize).


maxRequests?

optional maxRequests?: number;

Max concurrent tile requests per overlay layer (deck.gl maxRequests).


refinementStrategy?

optional refinementStrategy?: "best-available" | "no-overlap" | "never";

How deck.gl keeps/discards overscaled parent/child tiles (deck.gl refinementStrategy). "no-overlap" frees parent tiles once children cover the viewport, keeping the resident set smallest; "best-available" (deck default) keeps them to avoid flashes.

Last updated on