Migration, Hybrid, and Operational Best Practices
Domain 8 - Migration, Hybrid, and Operational Best Practices. Questions ask you to plan a move, connect it back to on-premises, and keep it running afterwards.
Migration approach
Inventory and dependency-map first, then choose a strategy per workload: rehost, replatform (managed database, containers), refactor, or retire. Pilot one workload, migrate in waves, then optimise sizing and cost. Data paths: Object Storage CLI/API for bulk files, logical dump or replication for databases, image-based rebuild for servers, and offline transfer when bandwidth makes the network path unrealistic.
Hybrid connectivity
- IPsec VPN: fast to provision, runs over the internet, latency varies.
- Dedicated/private connection: predictable latency and bandwidth, longer lead time and higher cost.
- Global Route Manager: DNS-level failover between on-premises and cloud endpoints.
- Keep ACG and firewall rules consistent at both ends, and check CIDR overlap before starting.
Operational practices
Least privilege, resource tagging, template-based provisioning, multi-zone deployment, tested backups and restore drills, alarms with runbooks, budget alerts, and a change window for risky work.
Worked example
On-premises MySQL to Cloud DB for MySQL. Reasoning: lower the DNS TTL a day ahead; create the managed instance in a private subnet; replicate from on-premises over the VPN; validate row counts and checksums; cut over inside a maintenance window; keep the source read-only for a week as rollback.
Traps
- Lift-and-shift without right-sizing, then paying for idle capacity.
- Underestimating transfer time for large datasets.
- Overlapping CIDR ranges that block VPN or peering.
- Having no rollback plan, or testing disaster recovery for the first time during an incident.