Recipes

End-to-end worked examples. Each recipe is a complete, runnable pattern — data fixture, transforms, chart interop — not a fragment. Skim the list to find the closest match to your use case, then steal.

Available recipes:

• CPU metrics dashboard — rolling averages, band charts, outlier detection
• Error-rate dashboard — aggregated counts bucketed by time, per-host breakdown
• Streaming dashboard — live-series ingest from a WebSocket-style source, throttled React rendering
• Incremental ±σ baselines — move a per-host rolling baseline onto the ingest path (partitionBy.rolling().collect()) so each render is a gather, not a re-walk (~16× at dashboard scale)
• Frontend telemetry reporting — collect high-frequency timing events, sample p95 latency to a backend every 30 s, display the live p95 in React