bench: per-chunk stream latency (CRTT/CDV) and replay-driven stream scenarios - #3393
Conversation
🦋 Changeset detectedLatest commit: d75dfb9 The changes in this PR will be included in the next version bump. This PR includes changesets to release 0 packagesWhen changesets are added to this PR, you'll see the packages that this PR includes changesets for and the associated semver types Not sure what this means? Click here to learn what changesets are. Click here if you're a maintainer who wants to add another changeset to this PR |
📊 Workflow Benchmarkscommit Backend:
Streams
📈 STSO distribution vs main (inline / queue-hop histograms)1020 steps (inline) Cumulative STSO time: main 194368ms → this run 177670ms (Δ -16698ms, -9%) 📈 CRTT drill-down (RTT distributions & profiles)No RTT over stream progress (avg per tenth of stream, bars scaled min→max): RTT by chunk size (avg per log size bin, ~160B → ~12KB serialized, bars scaled min→max): Delivery jitter over stream progress (avg positive CDV per tenth of stream, bars scaled min→max): ℹ️ Metric definitions & methodologyStreams: writer/reader sustained rates (steady window, 10% trimmed each side), first-chunk RTT (the stream-open path, before any buffering/backpressure), CRTT percentiles, and worst delivery stall (CDV max). Cells are medians across iterations; per-run values in the artifacts. No 🔴/🟢 marks until targets attach. The collapsed STSO distribution section above buckets every step gap, split inline (same warm process — pure framework overhead) vs queue-hop (fresh process — dispatch, reinit, replay). The collapsed CRTT drill-down: per-variant RTT histograms (fixed log bins, Best/P75/P90/P99 deltas compare against the most recent benchmark run on Metrics — TTFS: time to first step body (in-deployment start() → first step body) · Fan-out TTFS: fan-out time to first step (in-deployment start() → first of the parallel step bodies to complete) · Fan-out TTLS: fan-out time to last step (in-deployment start() → last of the parallel step bodies to complete, i.e. when the Promise.all resolves) · STSO: step-to-step overhead (gap between consecutive step bodies) · WO: workflow overhead (whole-run time outside step bodies, in-deployment anchored) · CRTT: chunk round-trip time (per-chunk write → read latency, one clock domain: deployment → stream backend → same deployment) · CDV: chunk delay variation / delivery jitter (inter-arrival gap minus inter-write gap per seq-adjacent pair; skew-free; the row is each run's MAX positive value, so one stall moves it) Scenarios — step: one trivial no-op step, no stream; no hooks, so the run stays in turbo mode (in-process fast path) · stream: one streaming step; no hooks, so the run stays in turbo mode (in-process fast path) · hook + stream: registers a hook before one step, which exits turbo mode (dispatch path) · 1020 steps: 1020 trivial sequential steps; STSO is measured between consecutive steps in the given step ranges, and WO is the whole-run overhead outside step bodies · Promise.all(100 steps): 100 trivial no-op steps started together in a single Promise.all; Fan-out TTFS is the first of them to complete and Fan-out TTLS the last, both from the in-deployment clientStart, so their gap is the spread the runtime adds across the fan-out · paced control (100/s, 60B): the control: 300 tiny (~60B) deltas metronome-paced at 100/s — zero workload structure, so it reads the transport floor and flush cadence, and disambiguates transport-wide vs workload-specific when a replay row moves · size sweep (100/s, 160B-12KB): same pacing as the control with deltas padded in rotation across seven log-spaced sizes (~160B–12KB) — rotation decouples size from stream position, so it isolates whether chunk size causes latency · replay gateway-gpt-5.4-nano-2000t (1x): raw provider SSE cadence captured at the AI gateway boundary (gpt-5.4-nano, the most popular gateway model; per-token deltas p50 208B = the modal production chunk size), replayed exactly as measured — the typical customer's workload; its CDV is the typical customer's real delivery jitter · replay eve-gpt-5.6-sol-2000t (1x): a captured eve turn (gpt-5.6-sol, the most-used demanding eve model; ~2000 output tokens = production p50 turn length) replayed exactly as measured — eve's envelope protocol re-ships the cumulative message so sizes ramp 142B→13KB; the demanding outlier tenant's reality · replay eve-gpt-5.6-sol-2000t (2x): the same eve capture at 2x — the headroom/stress row; real fast-tier models emit the same chunk sizes at proportionally higher rate, so time compression is a faithful speed model · first chunk (pooled): every run's seq-0 RTT pooled across all stream scenarios — the first chunk precedes any workload differentiation, so pooling samples one shared stream-open path with exact percentiles Replay cadences (semantic sha256) — eve-gpt-5.6-sol-2000t 🔴 marks a percentile over its target (within target is left unmarked). Targets (p75/p90/p99, ms) — TTFS 200/300/600 All timestamps are deployment-side; runs are triggered in-deployment, so the CI runner and api.vercel.com sit outside every measured window. TTFS = Cold starts stay in the numbers (real bursty-workload latency, inflates P75+); Best is the warm floor. |
🧪 E2E Test Results✅ All tests passed E2E Test SummarySummary
Details by Category✅ ▲ Vercel Production
✅ 💻 Local Development
✅ 📦 Local Production
✅ 🐘 Local Postgres
✅ 🪟 Windows
✅ vercel-multi-region
|
Sim WorldSimulated world deterministic testing for races. Traces 🟠 Mint-ordered log — 6 fail of 41 total
Full trace: 🟢 Append-only log — 0 fail of 41 total
Full trace: |
…cenarios Adds per-chunk stream latency to the CI benchmark report: CRTT (chunk round-trip time, same-clock write->read per chunk, aggregated inside the reader step into fixed log-bin histograms, index buckets, and mean-RTT profiles), CDV (chunk delay variation - skew-free delivery jitter, each run's max positive value), a pooled first-chunk RTT row (the stream-open path; exact percentiles), and replay scenarios driven by real captured cadences at the eve (envelope-protocol outlier) and AI-gateway (typical customer) boundaries - workload measured, not invented; the 2x speed multiplier on the eve stress row is the only chosen parameter. Stream scenarios render in their own table (rates, first-chunk RTT, CRTT percentiles, CDV max; medians across iterations, no targets yet). SL/SO report rows are retired: CRTT's seq-0 slice reproduces SL and its aggregate reproduces SO at ~100x the samples; write slip stays as artifact-only producer-stall data. Captures carry semantic sha256 hashes over canonical event tuples for cross-system identity with durabench. Signed-off-by: Alex Langenfeld <alex.langenfeld@vercel.com>
medianOf's even-count average produces values like 54.650000000000006, which rendered verbatim and blew out the stream table's column widths. Round the median to two decimals at the source (it lands in artifacts and history too) and round sub-100 values to one decimal in formatMs as the display-side guard. Signed-off-by: Alex Langenfeld <alex.langenfeld@vercel.com>
Signed-off-by: Alex Langenfeld <alex.langenfeld@vercel.com>
|
No backport to This commit adds new benchmark capabilities — CRTT/CDV per-chunk stream latency metrics, new bench workflows ( To override, re-run the Backport to stable workflow manually via |
2 participants
Summary & Motivation
(offsetMs, bytes)tuples so durabench's independent copy can be checked for drift.Test Plan