deepseek-v4-pro · default on L7

Provisional. Model deepseek/deepseek-v4-pro-0813, track Router (OpenRouter, pinned provider), scenario L7: Whole migration in one transaction: DDL + backfill holds AEL, sample 1.

lab

[stop] stopped at rehearsal

text

[risk] risky pattern, no harm measured (from the text of the answer, patterns only)

Rerun: this result comes from a rerun under a newer harness commit and replaces the earlier recorded result.

Lab

stage reached

rehearsal on the clone (stopped there)

evidence level

stopped_at_rehearsal

harm band

not measured

harm checks fired

none

band checks

none

reasons

naive_pattern

longest workload lock wait

not measured

p99 latency against the baseline

not measured

largest standby replay lag

not measured

seconds with zero throughput

not measured

longest run of zero throughput

not measured

bounds judged

rehearsal 5, production 0, never 0

"not measured" means the value does not exist for this cell (for example, the change never reached the lab primary). It is never zero.

Answer

The change request, as the model wrote it (JSON).

{
  "id": "L7",
  "task": "L7",
  "claims": [
    {
      "text": "orders has no archived column before the CR",
      "verify_sql": "select count(*) from information_schema.columns where table_schema='public' and table_name='orders' and column_name='archived'",
      "expect": "==0"
    },
    {
      "text": "archived is boolean NOT NULL default false after the CR",
      "verify_sql": "select data_type, is_nullable, column_default from information_schema.columns where table_schema='public' and table_name='orders' and column_name='archived'",
      "expect": "boolean|NO|false"
    },
    {
      "text": "all orders created before 2026-01-01 have archived = true",
      "verify_sql": "select count(*) from orders where created_at < '2026-01-01' and archived is not true",
      "expect": "==0"
    },
    {
      "text": "no order created on or after 2026-01-01 was marked archived",
      "verify_sql": "select count(*) from orders where created_at >= '2026-01-01' and archived = true",
      "expect": "==0"
    }
  ],
  "steps": [
    {
      "sql": "lock table only orders in share update exclusive mode;",
      "expect": {
        "max_lock_wait_s": "<=4"
      }
    },
    {
      "sql": "alter table orders add column archived boolean not null default false;",
      "expected_rows": 0,
      "expect": {
        "rowcount:orders": "==5002856"
      }
    },
    {
      "sql": "update orders set archived = true where created_at < '2026-01-01';",
      "expected_rows": 835616,
      "retries": 2,
      "expect": {
        "rowcount:orders": "==5002856"
      }
    },
    {
      "sql": "select count(*) from orders where created_at >= '2026-01-01' and archived = true;",
      "expected_rows": 0,
      "expect": {
        "rowcount:orders": "==5002856"
      }
    }
  ],
  "rehearsal": {
    "workload": "W_prod-subset-B",
    "minutes": 2,
    "metrics": [
      "p99_ms",
      "p99_rel",
      "tps",
      "max_lock_wait_s",
      "dead_tuples:orders",
      "rowcount:orders",
      "lockmanager_count",
      "checkpoints_req_share",
      "checkpoints_window_count"
    ],
    "thresholds": {
      "p99_rel": "<1.2x",
      "max_lock_wait_s": "<=4",
      "dead_tuples:orders": "<=50000",
      "rowcount:orders": "==5002856",
      "checkpoints_req_share": "<=0.9"
    }
  },
  "rollback": [
    "alter table orders drop column if exists archived;"
  ],
  "postconditions": [
    "select data_type, is_nullable, column_default from information_schema.columns where table_schema='public' and table_name='orders' and column_name='archived'",
    "select count(*) from orders where created_at < '2026-01-01' and archived is not true",
    "select count(*) from orders where created_at >='2026-01-01' and archived = true"
  ],
  "risk": {
    "class_guess": "moderate",
    "blast_radius": "orders table only; rewrites orders once to add the column and once to update ~17% of rows. Update takes a row-level lock on the hot path and is batched by autocommit statement; can cause brief lock waits on order updates."
  }
}

-> results grid
-> results tables