FleekDash 2.5.5 evolves the browser Service Worker path that already backs dashboard charts (first shipped for charts in 2.1.8). This release refines cache behavior specifically for chart-related requests: you spend less time waiting on the wire when you revisit the same windows, and the UI stays usable when the network is flaky, as long as that payload was fetched recently enough to be cached.
What's New
Stale-while-revalidate for chart payloads: after an analytics request succeeds, the Service Worker can return the last good JSON immediately from cache while refreshing from the network in the background, so the chart paints without an blank moment, then silently catches up when fresher data arrives.
Stronger repeat-session wins: flipping date presets, revisiting #/dashboard during the same admin session, or re-opening ranges you already explored hits warm cache paths more often, fewer duplicate round trips for the same logical slice.
Clearer separation from generic API traffic: chart-heavy endpoints follow dedicated cache rules tuned for short-lived analytics JSON, distinct from longer-lived static bundles, so chart caching stays aggressive where it helps without widening side effects across unrelated REST calls.
How it works (plain English)
On supported browsers, FleekDash registers a Service Worker scoped to the admin experience.
When dashboard charts request analytics data (typically REST/WP JSON), the worker applies a chart-oriented strategy: prefer showing cached responses quickly, then revalidate so numbers do not stay stale forever.
Cache keys derive from the effective request (including route parameters relevant to the chart window), so “same range, same query” tends to hit, while “new range” correctly misses and fetches fresh data.
Benefits you feel
Faster UI when operators tweak filters or bounce between dashboard and other screens, the chart layer does not start from scratch every time.
More resilient demos on slow Wi‑Fi or brief outages: recently fetched windows can still render from cache while connectivity recovers (within TTL limits, not a full offline analytics product).
Less redundant server work when many widgets ask for overlapping windows, clients reuse cached bodies instead of hammering identical endpoints.
What's Improved
Predictable dashboard rhythm: dense chart dashboards feel closer to a native app, less loading flicker when exploring date ranges you already loaded once this session.
