{
  "slug": "mongodb-aws-rerun",
  "campaign": "RERUN",
  "cloud": "aws",
  "agent": "mongodb",
  "status": "final",
  "note": "FINAL. The re-run pass: one fresh attempt per scenario that had failed in the reliability or coverage ledger, run after its failure class was fixed and loaded (26\u201331 Oct). Converged evidence, banked separately, never folded into the first-attempt ledgers. Gate v1 (4-part, write gating); 12b on its reachability-plus-write gate as its spec declares. Rows map to: 11#1 \u2192 reliability 11#5 (HARNESS); 24#1 \u2192 coverage 24#1 (PLATFORM); 07#1 \u2192 reliability 07#6 (PLATFORM); 12b#1 \u2192 reliability 12b#7 (PLATFORM); 20#1 \u2192 coverage 20#2/20#3 (AGENT); 21#1 \u2192 coverage 21#1\u20133 (PLATFORM + AGENT); 22#1 \u2192 coverage 22#1\u20133 (PLATFORM + AGENT). 07#1 failed on a NEW agent class \u2014 the CloudWatch agent started from a path only preinstalled images have, so logs never shipped on a healthy, writable database \u2014 a plan-stage rule was added and the slot run once more as 07#2: PASS on all four legs (31 Oct, 684 s deploy, 25 mongod log entries shipped from the first render). Off-ledger, each re-run: three rows (07, 12b, 20) first ran on 26 Oct with the write leg false under the defective probe and were re-run on 30 Oct under the corrected one; 21's first attempt was cut by a provider credit exhaustion (26 Oct); 22's first attempt failed on two defects of our own probe (strict verification of the operator's self-signed CA; TLS assumed on a render that declared it disabled), both fixed, then a third credit exhaustion (30 Oct), before the 31 Oct attempt passed.",
  "gate_v11_after": "2099-01-01T00:00:00Z",
  "source_file": "mongodb_aws_rerun_rows.jsonl",
  "extracted_at": "2026-10-30T23:24Z",
  "total": {
    "n": 8,
    "passed": 7
  },
  "attributed_failures": 1,
  "unattributed_failures": []
}