power-bi-performancelisted

# power-bi-performance Query tracing, benchmarking, VertiPaq analysis, and storage/formula engine diagnostics. ## Quick Reference ```bash # Trace a DAX query pbi trace start --query "EVALUATE SUMMARIZE(Sales, Calendar[Year], \"Rev\", SUM(Sales[Revenue]))" pbi trace start --measure "Total Revenue" --duration 30s pbi trace start --query-file ./queries/slow.dax --output ./traces/slow-trace.json # Benchmark a measure pbi benchmark --measure "Total Revenue" --iterations 10 pbi benchmark --measure "YTD Revenue" --iterations 20 --json # VertiPaq / model size analysis pbi trace model-stats pbi trace model-stats --json | jq '.tables | sort_by(.sizeBytes) | reverse | .[0:5]' ``` --- ## Worked Example 1: Profile a slow visual ```bash # 1 — Capture a trace while the slow visual refreshes pbi trace start --duration 60s --output ./traces/dashboard-trace.json # (manually refresh the slow visual in Desktop during the 60s window) # 2 — Inspect the trace cat ./traces/dashboard-trace.json | jq '.events[] | select(.durationMs > 500)' # 3 — Benchmark the suspected measure pbi benchmark --measure "Complex KPI" --iterations 20 --json ``` --- ## Worked Example 2: VertiPaq table size analysis ```bash pbi trace model-stats --json > model-stats.json # Top 5 tables by size jq '.tables | sort_by(.sizeBytes) | reverse | .[0:5] | .[] | {name: .name, sizeMB: (.sizeBytes / 1048576 | round), rows: .rowCount, cardinality: .columnCardinality}' model-stats.json ``` Expected outpu