python3 student-tests/concurrency.py --base-url "http://localhost:18080" --clients 10 --messages-per-client 10 --output evidence/concurrency.json 
{
  "lab": 3,
  "workload": "closed-loop parallel POST messages from independent clients",
  "clients": 10,
  "messages_per_client": 10,
  "expected_messages": 100,
  "acknowledged_201": 100,
  "stored_in_history": 100,
  "unique_stored_ids": 100,
  "unique_stored_texts": 100,
  "duplicates": 0,
  "missing_texts": [],
  "failed_writes": 0,
  "write_errors": [],
  "history_pages": 1,
  "history_order_strictly_increasing": true,
  "writers_elapsed_seconds": 0.455,
  "successful_rps": 219.78,
  "per_client": [
    {
      "client_index": 0,
      "acknowledged": 10,
      "errors": 0,
      "duration_seconds": 0.412
    },
    {
      "client_index": 1,
      "acknowledged": 10,
      "errors": 0,
      "duration_seconds": 0.41
    },
    {
      "client_index": 2,
      "acknowledged": 10,
      "errors": 0,
      "duration_seconds": 0.405
    },
    {
      "client_index": 3,
      "acknowledged": 10,
      "errors": 0,
      "duration_seconds": 0.401
    },
    {
      "client_index": 4,
      "acknowledged": 10,
      "errors": 0,
      "duration_seconds": 0.348
    },
    {
      "client_index": 5,
      "acknowledged": 10,
      "errors": 0,
      "duration_seconds": 0.269
    },
    {
      "client_index": 6,
      "acknowledged": 10,
      "errors": 0,
      "duration_seconds": 0.352
    },
    {
      "client_index": 7,
      "acknowledged": 10,
      "errors": 0,
      "duration_seconds": 0.338
    },
    {
      "client_index": 8,
      "acknowledged": 10,
      "errors": 0,
      "duration_seconds": 0.411
    },
    {
      "client_index": 9,
      "acknowledged": 10,
      "errors": 0,
      "duration_seconds": 0.403
    }
  ],
  "note": "Client-observed timings: closed-loop, single Docker host, no retries. Latency percentiles are intentionally absent here (see make load)."
}
PASS: 100 parallel messages from 10 clients stored exactly once
