diff --git a/docs/performance/runs/2026-09-23-north-star-steady-v1-harness-capacity-33c95bf/analysis.md b/docs/performance/runs/2026-09-23-north-star-steady-v1-harness-capacity-33c95bf/analysis.md
new file mode 100644
index 0000000..acf7644
--- /dev/null
+++ b/docs/performance/runs/2026-09-23-north-star-steady-v1-harness-capacity-33c95bf/analysis.md
@@ -0,0 +1,90 @@
+# Run Analysis
+
+Run: [run.json](run.json)
+
+Results: [result.json](result.json)
+
+This document is generated by AI from the run metadata, machine results, and
+retained diagnostics.
+
+## Purpose
+
+Locate the local capacity ceiling of the `kq-bench` harness on the
+`north-star-steady-v1` execution-time distribution (250 ms average, 475 ms p95,
+646 ms p99) by scaling producers, workers, worker concurrency, and ready
+partitions.
+
+## Observations
+
+Capacity probes at 3,000 RPS with 4 workers and 16 ready partitions:
+
+| Concurrency | Active worker RPS | Per worker | Drain | Queue p50 |
+| ---: | ---: | ---: | ---: | ---: |
+| 100 | 529.7 | 132 | 93.3 s | 44.6 s |
+| 250 | 1,136.2 | 284 | 32.8 s | 15.8 s |
+| 500 | 2,006.3 | 502 | 9.9 s | 3.5 s |
+| 1000 | 2,742.1 | 686 | 1.9 s | 501 ms |
+
+Arrival-rate scaling:
+
+| Requested RPS | Producers | Workers x concurrency | Partitions | Enqueue RPS | Worker RPS | Drain | Queue p50 / p95 / p99 | Host busy |
+| ---: | ---: | ---: | ---: | ---: | ---: | ---: | ---: | ---: |
+| 500 | 4 | 4 x 100 | 4 | 500.0 | 473.5 | 1.7 s | 438 / 982 / 1,295 ms | n/a |
+| 5,000 | 8 | 8 x 1000 | 16 | 4,995 | 4,613 | 1.7 s | 475 / 984 / 1,206 ms | 68% |
+| 8,000 | 10 | 12 x 1000 | 24 | 7,992 | 7,395 | 1.6 s | 475 / 997 / 1,211 ms | 87% |
+| 12,000 | 15 | 18 x 1000 | 32 | 11,984 | 11,114 | 1.6 s | 479 / 1,001 / 1,214 ms | 99% |
+| 16,000 | 20 | 24 x 1000 | 32 | 15,391 | 12,446 | 5.0 s | 700 / 4,087 / 5,233 ms | 100% |
+| 16,000 | 16 | 12 x 2000 | 32 | 15,958 | 12,760 | 5.1 s | 813 / 4,770 / 6,967 ms | 100% |
+| 20,000 | 20 | 15 x 2000 | 40 | 16,925 | 13,463 | 6.1 s | 1,025 / 7,962 / 10,367 ms | 100% |
+
+- 12,000 RPS was the highest sustainable tested rate; 16,000 RPS was the first
+  overloaded rate. Maximum observed worker throughput was 13,463 tasks/s.
+- At 16,000 and 20,000 RPS producers also fell behind the requested rate
+  (15,391 and 16,925 enqueues/s).
+- Steady-state CPU at 12,000 RPS, measured as cumulative process CPU time over
+  a 20 s window: Kafka Docker VM 4.29 cores, producers 2.75, harness Python
+  1.73, workers 0.77, other host processes 1.03.
+- Harness Python resident memory reached about 2.6 GB during the 45 s
+  12,000 RPS point.
+- All 2,815,000 measured tasks across all points completed with zero missing,
+  duplicate, outstanding, or dead-lettered IDs.
+- Measured execution averaged 250 to 252 ms with a 645 to 657 ms p99. Execution
+  maximum rose from about 1,494 ms to 1,805 ms once the host was saturated.
+
+## Interpretation
+
+- The ceiling is host CPU saturation, not a kq-specific limit. Workers used
+  0.77 cores for over 11,000 tasks/s; the harness's own Python accounting and
+  producer-side event emission consumed several cores of the same 12-core host.
+- Each harness producer calls `Enqueue` serially, so every task is an
+  individual `acks=all` produce request. Hypothesis: this drives a large share
+  of producer and Kafka CPU; the lean-probe run measures concurrent `Enqueue`
+  separately.
+- Queue p50 of about 475 ms is nearly independent of arrival rate at
+  concurrency 1000. It reflects worker generation behavior rather than backlog;
+  the queue-latency diagnosis run investigates it.
+- The rising execution maximum at saturated points suggests timer overshoot
+  from CPU contention.
+
+## Suggestions
+
+1. Rework the harness before using it above about 10,000 RPS: concurrent or
+   batched producer enqueues, compact or sampled per-task events, and
+   streaming accounting instead of in-memory rescans every second.
+2. Record per-process CPU and memory in the harness so saturated runs are
+   identified automatically.
+3. Compare against the lean-probe capacity run, which removes harness overhead
+   on the same host.
+
+## Caveats
+
+- Each point ran once on one arm64 laptop with a single Dockerized Kafka broker
+  (RF=1, 12 CPUs and 8 GB assigned to the Docker VM). Load generators, workers,
+  and Kafka shared the host.
+- Points used 20 to 45 s arrival windows; longer runs were not attempted
+  because harness memory grows with event count.
+- Host and process CPU peaks come from 2 s samples of `top`, `docker stats`,
+  and `ps`; `ps` CPU is a decaying average. The 12,000 RPS breakdown uses
+  cumulative CPU time and is more reliable.
+- The harness was invoked from source because the configured package index
+  rejected the `hatchling` build dependency; the harness code was unchanged.
diff --git a/docs/performance/runs/2026-09-23-north-star-steady-v1-harness-capacity-33c95bf/result.json b/docs/performance/runs/2026-09-23-north-star-steady-v1-harness-capacity-33c95bf/result.json
new file mode 100644
index 0000000..ffb6650
--- /dev/null
+++ b/docs/performance/runs/2026-09-23-north-star-steady-v1-harness-capacity-33c95bf/result.json
@@ -0,0 +1,556 @@
+{
+  "capacity": {
+    "highest_sustainable_tested_rps": 12000,
+    "first_overloaded_tested_rps": 16000,
+    "maximum_observed_active_worker_rps": 13462.589,
+    "maximum_observed_enqueue_rps": 16925.084
+  },
+  "cpu_breakdown_12k_rps": {
+    "source_point": "cpu12k",
+    "method": "per-process cumulative CPU time delta over a 20 s steady-state window, divided by wall time",
+    "cores": {
+      "kafka_docker_vm": 4.29,
+      "producers": 2.75,
+      "harness_python": 1.73,
+      "workers": 0.77,
+      "mover_and_observer": 0.0,
+      "other_host": 1.03
+    }
+  },
+  "correctness": {
+    "produced": 2815000,
+    "completed": 2815000,
+    "dead_lettered": 0,
+    "duplicates": 0,
+    "missing": 0,
+    "outstanding": 0
+  },
+  "points": [
+    {
+      "label": "smoke",
+      "requested_rps": 500,
+      "duration_seconds": 30,
+      "producers": 4,
+      "workers": 4,
+      "worker_concurrency": 100,
+      "ready_partitions": 4,
+      "achieved_enqueue_rps": 499.955,
+      "active_worker_rps": 473.519,
+      "drain_ms": 1699.773,
+      "sustained": true,
+      "queue_p50_ms": 437.519,
+      "queue_p95_ms": 982.153,
+      "queue_p99_ms": 1294.791,
+      "delivery_p50_ms": 688.988,
+      "delivery_p95_ms": 1280.133,
+      "delivery_p99_ms": 1579.452,
+      "enqueue_p50_ms": 0.689,
+      "enqueue_p99_ms": 2.77,
+      "execution_average_ms": 251.598,
+      "execution_p99_ms": 656.692,
+      "execution_max_ms": 1494.014,
+      "correctness": {
+        "completed": 15000,
+        "dead_lettered": 0,
+        "duplicates": 0,
+        "missing": 0,
+        "outstanding": 0,
+        "produced": 15000
+      },
+      "sampled_peaks": null,
+      "command": "kq-bench run --scenario north-star-steady-v1 --profile ci-smoke --output <tmp>/kqperf --set workload.arrival.target_rps=500 --set workload.arrival.duration_seconds=30 --set topology.producers=4 --set topology.workers=4 --set topology.worker_concurrency=100 --set topology.ready_partitions=4"
+    },
+    {
+      "label": "capC100",
+      "requested_rps": 3000,
+      "duration_seconds": 20,
+      "producers": 6,
+      "workers": 4,
+      "worker_concurrency": 100,
+      "ready_partitions": 16,
+      "achieved_enqueue_rps": 2998.839,
+      "active_worker_rps": 529.745,
+      "drain_ms": 93277.057,
+      "sustained": false,
+      "queue_p50_ms": 44575.481,
+      "queue_p95_ms": 86987.014,
+      "queue_p99_ms": 90402.088,
+      "delivery_p50_ms": 44840.198,
+      "delivery_p95_ms": 87243.293,
+      "delivery_p99_ms": 90627.816,
+      "enqueue_p50_ms": 0.315,
+      "enqueue_p99_ms": 0.944,
+      "execution_average_ms": 250.68,
+      "execution_p99_ms": 650.362,
+      "execution_max_ms": 1493.971,
+      "correctness": {
+        "completed": 60000,
+        "dead_lettered": 0,
+        "duplicates": 0,
+        "missing": 0,
+        "outstanding": 0,
+        "produced": 60000
+      },
+      "sampled_peaks": {
+        "host_busy_percent": 50.6,
+        "kafka_container_cpu_percent": 216.7,
+        "kq_load_cpu_percent": 58,
+        "kq_load_rss_mb": 254,
+        "harness_python_cpu_percent": 29,
+        "harness_python_rss_mb": 270
+      },
+      "command": "kq-bench run --scenario north-star-steady-v1 --profile ci-smoke --output <tmp>/capC100 --set workload.arrival.target_rps=3000 --set workload.arrival.duration_seconds=20 --set topology.producers=6 --set topology.workers=4 --set topology.worker_concurrency=100 --set topology.ready_partitions=16 --set timeouts.drain_seconds=240"
+    },
+    {
+      "label": "capC250",
+      "requested_rps": 3000,
+      "duration_seconds": 20,
+      "producers": 6,
+      "workers": 4,
+      "worker_concurrency": 250,
+      "ready_partitions": 16,
+      "achieved_enqueue_rps": 2998.906,
+      "active_worker_rps": 1136.216,
+      "drain_ms": 32829.499,
+      "sustained": false,
+      "queue_p50_ms": 15778.688,
+      "queue_p95_ms": 30174.639,
+      "queue_p99_ms": 31711.795,
+      "delivery_p50_ms": 16030.769,
+      "delivery_p95_ms": 30418.353,
+      "delivery_p99_ms": 32002.521,
+      "enqueue_p50_ms": 0.36,
+      "enqueue_p99_ms": 1.054,
+      "execution_average_ms": 250.526,
+      "execution_p99_ms": 650.691,
+      "execution_max_ms": 1493.636,
+      "correctness": {
+        "completed": 60000,
+        "dead_lettered": 0,
+        "duplicates": 0,
+        "missing": 0,
+        "outstanding": 0,
+        "produced": 60000
+      },
+      "sampled_peaks": {
+        "host_busy_percent": 69.2,
+        "kafka_container_cpu_percent": 316.5,
+        "kq_load_cpu_percent": 70,
+        "kq_load_rss_mb": 260,
+        "harness_python_cpu_percent": 24,
+        "harness_python_rss_mb": 269
+      },
+      "command": "kq-bench run --scenario north-star-steady-v1 --profile ci-smoke --output <tmp>/capC250 --set workload.arrival.target_rps=3000 --set workload.arrival.duration_seconds=20 --set topology.producers=6 --set topology.workers=4 --set topology.worker_concurrency=250 --set topology.ready_partitions=16 --set timeouts.drain_seconds=240"
+    },
+    {
+      "label": "capC500",
+      "requested_rps": 3000,
+      "duration_seconds": 20,
+      "producers": 6,
+      "workers": 4,
+      "worker_concurrency": 500,
+      "ready_partitions": 16,
+      "achieved_enqueue_rps": 2999.941,
+      "active_worker_rps": 2006.263,
+      "drain_ms": 9934.554,
+      "sustained": false,
+      "queue_p50_ms": 3508.021,
+      "queue_p95_ms": 10833.22,
+      "queue_p99_ms": 12374.796,
+      "delivery_p50_ms": 3762.021,
+      "delivery_p95_ms": 11081.333,
+      "delivery_p99_ms": 12631.386,
+      "enqueue_p50_ms": 0.291,
+      "enqueue_p99_ms": 0.741,
+      "execution_average_ms": 250.415,
+      "execution_p99_ms": 650.226,
+      "execution_max_ms": 1493.578,
+      "correctness": {
+        "completed": 60000,
+        "dead_lettered": 0,
+        "duplicates": 0,
+        "missing": 0,
+        "outstanding": 0,
+        "produced": 60000
+      },
+      "sampled_peaks": {
+        "host_busy_percent": 75.0,
+        "kafka_container_cpu_percent": 260.68,
+        "kq_load_cpu_percent": 59,
+        "kq_load_rss_mb": 266,
+        "harness_python_cpu_percent": 38,
+        "harness_python_rss_mb": 263
+      },
+      "command": "kq-bench run --scenario north-star-steady-v1 --profile ci-smoke --output <tmp>/capC500 --set workload.arrival.target_rps=3000 --set workload.arrival.duration_seconds=20 --set topology.producers=6 --set topology.workers=4 --set topology.worker_concurrency=500 --set topology.ready_partitions=16 --set timeouts.drain_seconds=240"
+    },
+    {
+      "label": "capC1000",
+      "requested_rps": 3000,
+      "duration_seconds": 20,
+      "producers": 6,
+      "workers": 4,
+      "worker_concurrency": 1000,
+      "ready_partitions": 16,
+      "achieved_enqueue_rps": 2998.195,
+      "active_worker_rps": 2742.143,
+      "drain_ms": 1896.079,
+      "sustained": true,
+      "queue_p50_ms": 500.606,
+      "queue_p95_ms": 1072.521,
+      "queue_p99_ms": 1312.133,
+      "delivery_p50_ms": 748.943,
+      "delivery_p95_ms": 1367.854,
+      "delivery_p99_ms": 1634.258,
+      "enqueue_p50_ms": 0.369,
+      "enqueue_p99_ms": 1.397,
+      "execution_average_ms": 250.387,
+      "execution_p99_ms": 650.482,
+      "execution_max_ms": 1494.019,
+      "correctness": {
+        "completed": 60000,
+        "dead_lettered": 0,
+        "duplicates": 0,
+        "missing": 0,
+        "outstanding": 0,
+        "produced": 60000
+      },
+      "sampled_peaks": {
+        "host_busy_percent": 54.7,
+        "kafka_container_cpu_percent": 286.67,
+        "kq_load_cpu_percent": 68,
+        "kq_load_rss_mb": 281,
+        "harness_python_cpu_percent": 97,
+        "harness_python_rss_mb": 266
+      },
+      "command": "kq-bench run --scenario north-star-steady-v1 --profile ci-smoke --output <tmp>/capC1000 --set workload.arrival.target_rps=3000 --set workload.arrival.duration_seconds=20 --set topology.producers=6 --set topology.workers=4 --set topology.worker_concurrency=1000 --set topology.ready_partitions=16 --set timeouts.drain_seconds=240"
+    },
+    {
+      "label": "r5k",
+      "requested_rps": 5000,
+      "duration_seconds": 20,
+      "producers": 8,
+      "workers": 8,
+      "worker_concurrency": 1000,
+      "ready_partitions": 16,
+      "achieved_enqueue_rps": 4994.773,
+      "active_worker_rps": 4613.395,
+      "drain_ms": 1684.682,
+      "sustained": true,
+      "queue_p50_ms": 475.407,
+      "queue_p95_ms": 984.087,
+      "queue_p99_ms": 1205.72,
+      "delivery_p50_ms": 724.801,
+      "delivery_p95_ms": 1284.392,
+      "delivery_p99_ms": 1520.589,
+      "enqueue_p50_ms": 0.337,
+      "enqueue_p99_ms": 1.082,
+      "execution_average_ms": 250.011,
+      "execution_p99_ms": 646.771,
+      "execution_max_ms": 1493.158,
+      "correctness": {
+        "completed": 100000,
+        "dead_lettered": 0,
+        "duplicates": 0,
+        "missing": 0,
+        "outstanding": 0,
+        "produced": 100000
+      },
+      "sampled_peaks": {
+        "host_busy_percent": 67.6,
+        "kafka_container_cpu_percent": 189.5,
+        "kq_load_cpu_percent": 121,
+        "kq_load_rss_mb": 442,
+        "harness_python_cpu_percent": 53,
+        "harness_python_rss_mb": 393
+      },
+      "command": "kq-bench run --scenario north-star-steady-v1 --profile ci-smoke --output <tmp>/r5k --set workload.arrival.target_rps=5000 --set workload.arrival.duration_seconds=20 --set topology.producers=8 --set topology.workers=8 --set topology.worker_concurrency=1000 --set topology.ready_partitions=16 --set timeouts.drain_seconds=240"
+    },
+    {
+      "label": "r8k",
+      "requested_rps": 8000,
+      "duration_seconds": 20,
+      "producers": 10,
+      "workers": 12,
+      "worker_concurrency": 1000,
+      "ready_partitions": 24,
+      "achieved_enqueue_rps": 7991.509,
+      "active_worker_rps": 7394.717,
+      "drain_ms": 1643.775,
+      "sustained": true,
+      "queue_p50_ms": 474.791,
+      "queue_p95_ms": 997.449,
+      "queue_p99_ms": 1210.845,
+      "delivery_p50_ms": 724.336,
+      "delivery_p95_ms": 1293.945,
+      "delivery_p99_ms": 1528.183,
+      "enqueue_p50_ms": 0.42,
+      "enqueue_p99_ms": 1.589,
+      "execution_average_ms": 249.801,
+      "execution_p99_ms": 646.675,
+      "execution_max_ms": 1493.748,
+      "correctness": {
+        "completed": 160000,
+        "dead_lettered": 0,
+        "duplicates": 0,
+        "missing": 0,
+        "outstanding": 0,
+        "produced": 160000
+      },
+      "sampled_peaks": {
+        "host_busy_percent": 86.7,
+        "kafka_container_cpu_percent": 238.12,
+        "kq_load_cpu_percent": 232,
+        "kq_load_rss_mb": 602,
+        "harness_python_cpu_percent": 105,
+        "harness_python_rss_mb": 771
+      },
+      "command": "kq-bench run --scenario north-star-steady-v1 --profile ci-smoke --output <tmp>/r8k --set workload.arrival.target_rps=8000 --set workload.arrival.duration_seconds=20 --set topology.producers=10 --set topology.workers=12 --set topology.worker_concurrency=1000 --set topology.ready_partitions=24 --set timeouts.drain_seconds=240"
+    },
+    {
+      "label": "r12k",
+      "requested_rps": 12000,
+      "duration_seconds": 20,
+      "producers": 15,
+      "workers": 18,
+      "worker_concurrency": 1000,
+      "ready_partitions": 32,
+      "achieved_enqueue_rps": 11984.156,
+      "active_worker_rps": 11114.172,
+      "drain_ms": 1593.796,
+      "sustained": true,
+      "queue_p50_ms": 479.039,
+      "queue_p95_ms": 1001.339,
+      "queue_p99_ms": 1213.749,
+      "delivery_p50_ms": 727.464,
+      "delivery_p95_ms": 1295.048,
+      "delivery_p99_ms": 1530.284,
+      "enqueue_p50_ms": 0.7,
+      "enqueue_p99_ms": 1.912,
+      "execution_average_ms": 249.566,
+      "execution_p99_ms": 644.815,
+      "execution_max_ms": 1496.496,
+      "correctness": {
+        "completed": 240000,
+        "dead_lettered": 0,
+        "duplicates": 0,
+        "missing": 0,
+        "outstanding": 0,
+        "produced": 240000
+      },
+      "sampled_peaks": {
+        "host_busy_percent": 99.1,
+        "kafka_container_cpu_percent": 330.02,
+        "kq_load_cpu_percent": 428,
+        "kq_load_rss_mb": 890,
+        "harness_python_cpu_percent": 192,
+        "harness_python_rss_mb": 1138
+      },
+      "command": "kq-bench run --scenario north-star-steady-v1 --profile ci-smoke --output <tmp>/r12k --set workload.arrival.target_rps=12000 --set workload.arrival.duration_seconds=20 --set topology.producers=15 --set topology.workers=18 --set topology.worker_concurrency=1000 --set topology.ready_partitions=32 --set timeouts.drain_seconds=240"
+    },
+    {
+      "label": "prof12k",
+      "requested_rps": 12000,
+      "duration_seconds": 40,
+      "producers": 15,
+      "workers": 18,
+      "worker_concurrency": 1000,
+      "ready_partitions": 32,
+      "achieved_enqueue_rps": 11983.926,
+      "active_worker_rps": 11384.692,
+      "drain_ms": 2143.885,
+      "sustained": true,
+      "queue_p50_ms": 477.783,
+      "queue_p95_ms": 1013.138,
+      "queue_p99_ms": 1293.222,
+      "delivery_p50_ms": 726.339,
+      "delivery_p95_ms": 1310.068,
+      "delivery_p99_ms": 1604.094,
+      "enqueue_p50_ms": 0.678,
+      "enqueue_p99_ms": 1.998,
+      "execution_average_ms": 249.5,
+      "execution_p99_ms": 644.864,
+      "execution_max_ms": 1804.656,
+      "correctness": {
+        "completed": 480000,
+        "dead_lettered": 0,
+        "duplicates": 0,
+        "missing": 0,
+        "outstanding": 0,
+        "produced": 480000
+      },
+      "sampled_peaks": {
+        "host_busy_percent": 100.0,
+        "kafka_container_cpu_percent": 287.55,
+        "kq_load_cpu_percent": 313,
+        "kq_load_rss_mb": 904,
+        "harness_python_cpu_percent": 184,
+        "harness_python_rss_mb": 2368
+      },
+      "command": "kq-bench run --scenario north-star-steady-v1 --profile ci-smoke --output <tmp>/prof12k --set workload.arrival.target_rps=12000 --set workload.arrival.duration_seconds=40 --set topology.producers=15 --set topology.workers=18 --set topology.worker_concurrency=1000 --set topology.ready_partitions=32 --set timeouts.drain_seconds=240"
+    },
+    {
+      "label": "cpu12k",
+      "requested_rps": 12000,
+      "duration_seconds": 45,
+      "producers": 15,
+      "workers": 18,
+      "worker_concurrency": 1000,
+      "ready_partitions": 32,
+      "achieved_enqueue_rps": 11991.237,
+      "active_worker_rps": 11535.614,
+      "drain_ms": 1807.56,
+      "sustained": true,
+      "queue_p50_ms": 485.252,
+      "queue_p95_ms": 1029.463,
+      "queue_p99_ms": 1314.801,
+      "delivery_p50_ms": 734.58,
+      "delivery_p95_ms": 1327.889,
+      "delivery_p99_ms": 1620.303,
+      "enqueue_p50_ms": 0.714,
+      "enqueue_p99_ms": 2.22,
+      "execution_average_ms": 249.505,
+      "execution_p99_ms": 644.966,
+      "execution_max_ms": 1804.004,
+      "correctness": {
+        "completed": 540000,
+        "dead_lettered": 0,
+        "duplicates": 0,
+        "missing": 0,
+        "outstanding": 0,
+        "produced": 540000
+      },
+      "sampled_peaks": {
+        "host_busy_percent": 100.0,
+        "kafka_container_cpu_percent": 340.58,
+        "kq_load_cpu_percent": 379,
+        "kq_load_rss_mb": 912,
+        "harness_python_cpu_percent": 196,
+        "harness_python_rss_mb": 2615
+      },
+      "command": "kq-bench run --scenario north-star-steady-v1 --profile ci-smoke --output <tmp>/cpu12k --set workload.arrival.target_rps=12000 --set workload.arrival.duration_seconds=45 --set topology.producers=15 --set topology.workers=18 --set topology.worker_concurrency=1000 --set topology.ready_partitions=32 --set timeouts.drain_seconds=240"
+    },
+    {
+      "label": "r16k",
+      "requested_rps": 16000,
+      "duration_seconds": 20,
+      "producers": 20,
+      "workers": 24,
+      "worker_concurrency": 1000,
+      "ready_partitions": 32,
+      "achieved_enqueue_rps": 15391.477,
+      "active_worker_rps": 12446.131,
+      "drain_ms": 4951.005,
+      "sustained": false,
+      "queue_p50_ms": 699.733,
+      "queue_p95_ms": 4086.928,
+      "queue_p99_ms": 5233.334,
+      "delivery_p50_ms": 966.663,
+      "delivery_p95_ms": 4397.416,
+      "delivery_p99_ms": 5521.904,
+      "enqueue_p50_ms": 0.882,
+      "enqueue_p99_ms": 2.623,
+      "execution_average_ms": 249.598,
+      "execution_p99_ms": 645.282,
+      "execution_max_ms": 1594.017,
+      "correctness": {
+        "completed": 320000,
+        "dead_lettered": 0,
+        "duplicates": 0,
+        "missing": 0,
+        "outstanding": 0,
+        "produced": 320000
+      },
+      "sampled_peaks": {
+        "host_busy_percent": 100.0,
+        "kafka_container_cpu_percent": 322.76,
+        "kq_load_cpu_percent": 457,
+        "kq_load_rss_mb": 1171,
+        "harness_python_cpu_percent": 180,
+        "harness_python_rss_mb": 1496
+      },
+      "command": "kq-bench run --scenario north-star-steady-v1 --profile ci-smoke --output <tmp>/r16k --set workload.arrival.target_rps=16000 --set workload.arrival.duration_seconds=20 --set topology.producers=20 --set topology.workers=24 --set topology.worker_concurrency=1000 --set topology.ready_partitions=32 --set timeouts.drain_seconds=240"
+    },
+    {
+      "label": "r16k-c2000",
+      "requested_rps": 16000,
+      "duration_seconds": 20,
+      "producers": 16,
+      "workers": 12,
+      "worker_concurrency": 2000,
+      "ready_partitions": 32,
+      "achieved_enqueue_rps": 15957.66,
+      "active_worker_rps": 12759.856,
+      "drain_ms": 5055.301,
+      "sustained": false,
+      "queue_p50_ms": 812.852,
+      "queue_p95_ms": 4770.124,
+      "queue_p99_ms": 6967.437,
+      "delivery_p50_ms": 1176.959,
+      "delivery_p95_ms": 5077.714,
+      "delivery_p99_ms": 7389.238,
+      "enqueue_p50_ms": 0.811,
+      "enqueue_p99_ms": 2.006,
+      "execution_average_ms": 249.593,
+      "execution_p99_ms": 645.302,
+      "execution_max_ms": 1594.749,
+      "correctness": {
+        "completed": 320000,
+        "dead_lettered": 0,
+        "duplicates": 0,
+        "missing": 0,
+        "outstanding": 0,
+        "produced": 320000
+      },
+      "sampled_peaks": {
+        "host_busy_percent": 100.0,
+        "kafka_container_cpu_percent": 342.58,
+        "kq_load_cpu_percent": 355,
+        "kq_load_rss_mb": 846,
+        "harness_python_cpu_percent": 186,
+        "harness_python_rss_mb": 1569
+      },
+      "command": "kq-bench run --scenario north-star-steady-v1 --profile ci-smoke --output <tmp>/r16k-c2000 --set workload.arrival.target_rps=16000 --set workload.arrival.duration_seconds=20 --set topology.producers=16 --set topology.workers=12 --set topology.worker_concurrency=2000 --set topology.ready_partitions=32 --set timeouts.drain_seconds=240"
+    },
+    {
+      "label": "r20k-c2000",
+      "requested_rps": 20000,
+      "duration_seconds": 20,
+      "producers": 20,
+      "workers": 15,
+      "worker_concurrency": 2000,
+      "ready_partitions": 40,
+      "achieved_enqueue_rps": 16925.084,
+      "active_worker_rps": 13462.589,
+      "drain_ms": 6102.466,
+      "sustained": false,
+      "queue_p50_ms": 1025.412,
+      "queue_p95_ms": 7961.55,
+      "queue_p99_ms": 10366.812,
+      "delivery_p50_ms": 1505.95,
+      "delivery_p95_ms": 8295.118,
+      "delivery_p99_ms": 10695.774,
+      "enqueue_p50_ms": 0.991,
+      "enqueue_p99_ms": 2.837,
+      "execution_average_ms": 249.539,
+      "execution_p99_ms": 644.287,
+      "execution_max_ms": 1791.135,
+      "correctness": {
+        "completed": 400000,
+        "dead_lettered": 0,
+        "duplicates": 0,
+        "missing": 0,
+        "outstanding": 0,
+        "produced": 400000
+      },
+      "sampled_peaks": {
+        "host_busy_percent": 100.0,
+        "kafka_container_cpu_percent": 347.37,
+        "kq_load_cpu_percent": 367,
+        "kq_load_rss_mb": 1043,
+        "harness_python_cpu_percent": 182,
+        "harness_python_rss_mb": 2003
+      },
+      "command": "kq-bench run --scenario north-star-steady-v1 --profile ci-smoke --output <tmp>/r20k-c2000 --set workload.arrival.target_rps=20000 --set workload.arrival.duration_seconds=20 --set topology.producers=20 --set topology.workers=15 --set topology.worker_concurrency=2000 --set topology.ready_partitions=40 --set timeouts.drain_seconds=240"
+    }
+  ]
+}
diff --git a/docs/performance/runs/2026-09-23-north-star-steady-v1-harness-capacity-33c95bf/run.json b/docs/performance/runs/2026-09-23-north-star-steady-v1-harness-capacity-33c95bf/run.json
new file mode 100644
index 0000000..17d6b5b
--- /dev/null
+++ b/docs/performance/runs/2026-09-23-north-star-steady-v1-harness-capacity-33c95bf/run.json
@@ -0,0 +1,95 @@
+{
+  "run_id": "2026-09-23-north-star-steady-v1-harness-capacity-33c95bf",
+  "created_at": "2026-09-23T00:05:00Z",
+  "revision": "33c95bf74b5c",
+  "purpose": "Locate the local capacity ceiling of the kq-bench harness on the north-star execution-time distribution by scaling producers, workers, worker concurrency, and ready partitions.",
+  "scenario": {
+    "id": "north-star-steady-v1",
+    "version": 1
+  },
+  "profile": "ci-smoke",
+  "command": "kq-bench run --scenario north-star-steady-v1 --profile ci-smoke --set <per-point topology and arrival overrides>; see result.json points[].command",
+  "environment": {
+    "architecture": "arm64",
+    "ci_provider": "",
+    "cpu_count": 12,
+    "go_version": "go version go1.27.0 darwin/arm64",
+    "kafka_version": "apache/kafka:4.3.1",
+    "memory_bytes": 22118400,
+    "operating_system": "Darwin",
+    "python_version": "3.13.12",
+    "runner": "rays-MacBook-Pro.local",
+    "docker_vm_cpus": 12,
+    "docker_vm_memory_bytes": 8216776704
+  },
+  "topology": {
+    "ready_partitions_sweep": [
+      4,
+      16,
+      24,
+      32,
+      40
+    ],
+    "retry_partitions": 4,
+    "producers_sweep": [
+      4,
+      6,
+      8,
+      10,
+      15,
+      16,
+      20
+    ],
+    "workers_sweep": [
+      4,
+      8,
+      12,
+      15,
+      18,
+      24
+    ],
+    "worker_concurrency_sweep": [
+      100,
+      250,
+      500,
+      1000,
+      2000
+    ],
+    "movers": 1
+  },
+  "workload": {
+    "arrival": {
+      "target_rps_sweep": [
+        500,
+        3000,
+        5000,
+        8000,
+        12000,
+        16000,
+        20000
+      ],
+      "duration_seconds_sweep": [
+        20,
+        30,
+        40,
+        45
+      ]
+    },
+    "execution_time_ms": {
+      "average": 250,
+      "p95": 475,
+      "p99": 646
+    },
+    "random_seed": 1,
+    "failures": {
+      "rate": 0,
+      "mode": "none",
+      "duration_ms": 0
+    },
+    "warmup": {
+      "tasks": 10,
+      "id_base": 4294967296
+    }
+  },
+  "notes": "Every point used revision 33c95bf74b5c with no source changes. The harness was invoked from source (PYTHONPATH=bench) because the configured package index rejected the hatchling build dependency. Host and process CPU peaks were sampled every 2 s with top, docker stats, and ps; ps CPU is a decaying average and is approximate."
+}