{
  "name": "Warm dashboard reads through a four-proxy fleet",
  "recorded_on": "2026-10-08",
  "commit": "0f3cde7",
  "workload": {
    "queries": "Six per-customer dashboard lookups over 30 active customers from a 1,000-customer demo dataset",
    "protocol": "Postgres extended protocol, pgbench",
    "duration_seconds": 20,
    "concurrent_clients": 64,
    "writer": "quiet",
    "cache": "warm result caches"
  },
  "setup": {
    "machine": "AWS m7g.8xlarge, 32 cores",
    "proxies": 4,
    "cores_per_proxy": 2,
    "writer_and_reader_cores": 6,
    "load_balancer": "HAProxy in TCP mode, leastconn",
    "load_balancer_assigned_cores": 4,
    "client_assigned_cores": 14,
    "state_storage": "S3"
  },
  "results": {
    "reads_per_second": 3125.1,
    "median_response_ms": 20.37,
    "p99_response_ms": 40.51,
    "measured_responses": 52240
  },
  "correctness": {
    "queries_compared": 186,
    "answers_agree": true,
    "fleet_http_answers_agree": true,
    "fleet_postgres_answers_agree": true
  },
  "scope": "A recorded warm-cache dashboard lookup benchmark, not a measurement of arbitrary queries, exports, active-write freshness or subscription pricing.",
  "repository_sources": [
    "docs/results/bench/proxy/read_set_fleet_hot.json",
    "docs/findings/wire.md",
    "scripts/proxy_fleet.py",
    "scripts/snowflake/dashboard.sql"
  ]
}