Indexing and consistency

TL;DR

• Indexing keeps Firestore vault documents aligned with on-chain vault state.
• Most important actions trigger indexing automatically.
• If indexing fails, the app blocks with a Retry indexing modal until it succeeds.

When indexing runs

• After creating a vault.
• After opening/canceling/accepting a request.
• After repayment and claim processing.
• During manual retry flows.

What indexing does

1. Reads fresh vault state from chain.
2. Transforms state to the app document shape.
3. Writes back to Firestore for realtime subscribers.

Route behavior

• /api/index_vault performs immediate indexing and returns success/failure.
• /api/indexing/enqueue stores a background job for worker recovery.
• /api/indexing/worker processes queued jobs with retry backoff.

Symptoms of stale data

• Request state does not change after a confirmed transaction.
• Lender/owner action buttons do not match expected role.
• Discover and vault page show different state briefly.

Recovery steps

1. Wait a few seconds after wallet confirmation.
2. Use retry indexing where available.
3. Refresh the page.
4. If still stale, escalate with transaction hash.