Create & anchor records

Creating records

A record is a JSON payload under a domain, with a source:

curl -s -X POST $BASE/api/v1/audit/records -H 'Content-Type: application/json' -d '{
  "source": "payments",
  "payload": { "event": "charge", "amount": 1200 }
}'

The API computes the integrity hash and stores the record (after a durable WAL write). Records are append-only — a DELETE is a soft-delete that preserves the audit trail.

Controlling the hash

By default the whole payload feeds the hash. Use hash_fields to pin integrity to specific keys, so unrelated fields can change without breaking it:

{ "source": "crm", "payload": { "party": "acme", "amount": 100, "note": "draft" },
  "hash_fields": ["party", "amount"] }

Anchoring

Pending records are folded into a Merkle tree and the root is anchored on Fabric.

On demand

curl -s -X POST $BASE/api/v1/audit/records/batch
# -> { "batch_id": "...", "merkle_root": "...", "tx_id": "...", "anchored": true }

Automatically

Enable auto-batching so the processor anchors on a size or time trigger:

batching:
  enabled: true
  auto_batch_size: 100
  auto_batch_interval: 30s

Durability & correctness

• Zero loss: a record is fsync-ed to the WAL before it is acknowledged; a crash is recovered by idempotent replay.
• No double-anchoring: batch claiming is atomic, so concurrent batchers (or replicas) never anchor the same record twice.
• Self-healing: if anchoring fails, records are marked anchor_failed and a reconciler re-drives them — never silently lost.

Next: verify integrity.