Deployment Guide

The exact commands vary by stack, but a reliable rollout follows a controlled sequence. Do not expose a chain to users before the settlement, data availability, bridge, monitoring, and recovery paths have been verified together. Optimism's batcher setup guide shows the separate sequencer and batcher setup that a real rollout must coordinate.

1. Build and record immutable images, package versions, contract artifacts, and the reviewed configuration manifest.
2. Provision isolated environments for execution, databases, batch submission, proving or challenge services, and observability.
3. Deploy and verify settlement and bridge contracts using the intended governance and ownership model.
4. Start the core services with access to only the keys and endpoints each service needs.
5. Submit representative transactions, batches, and proofs or challenges through the complete settlement path.
6. Verify data retrieval, indexing, deposits, withdrawals, monitoring, alert delivery, backups, and a recovery drill.
7. Publish endpoints and onboarding instructions only after the acceptance criteria are met.

Keep a launch record containing the artifact digests, contract addresses, chain identity, configuration manifest, owners, health evidence, and rollback decision points.

Production gate

Use a named production gate rather than treating a successful deployment job as approval. The gate should confirm:

• Settlement transactions are accepted and confirmed at the expected cadence.
• Published batch data can be retrieved and used to reconstruct a sampled state range.
• Deposits and withdrawals complete through the intended bridge path.
• A sequencer or worker replacement does not lose state or create duplicate submissions.
• Dashboards receive all critical component signals and alerts reach the expected destination.
• Governance and emergency procedures have been exercised with the actual contract addresses and access controls.

First-week operating posture

Launch with heightened observation. Review batch age, settlement cost, proof or challenge performance, bridge completion time, data retrieval success, and error budgets daily. Keep changes small and reversible during this period. A new chain's normal operating baseline is learned from real traffic, so early changes need more evidence, not less.