MySQL 8.0 Cluster Deployment

I’m retiring our old Galera ring and rebuilding everything—this time on MySQL 8.0 with InnoDB Cluster—across seven physical hosts that currently peak at about 15 QPS each. The workload is balanced between reads and writes, so the end-state must support smooth, low-latency traffic distribution in both directions while giving us cleaner fail-over and simpler maintenance.

Here’s what needs to happen:

• Prepare each node: OS tuning, MySQL 8.0 install, secure configuration, and SSL.
• Create and initialise the InnoDB Cluster, wire up MySQL Router on every application tier, and verify automatic fail-over.
• Design and document the read/write split strategy so load is shared evenly across the seven hosts.
• Build monitoring hooks for replication lag, node health, and cluster status; integrate with our existing Prometheus stack if possible.
• Deliver a repeatable build script or Ansible playbook plus a concise topology diagram and recovery checklist.

Acceptance criteria
1. All seven nodes join the PRIMARY cluster and stay in sync under a simulated 20 QPS mixed workload.
2. A deliberate node failure triggers automatic re-routing within 15 seconds without losing writes.
3. The hand-over pack is clear enough for a new DBA to rebuild any node from scratch.

Once everything holds up under load-test, the project is complete.