
TL;DR
Switching to the target environment marks the start of ongoing operations. Monitoring, backups, recovery, maintenance and ownership therefore belong in the migration plan. A sound handover demonstrates that alerts reach the right people, data can be restored and the operations team can act when something fails. The provider's responsibilities and those retained by the customer must be explicitly agreed.
Monday morning after the move
The migration window is over. Applications run in the target environment and users start signing in. Then an order gets stuck during processing.
Who notices? Who investigates? And who informs the business team?
These questions determine whether the new cloud works in daily use. The target platform is one part of the change. People, processes and technical tools also need to work together.
1. Agree operational ownership before go-live
“The provider runs the cloud” is too vague as a task description. It might refer to hardware, virtualisation, Kubernetes or specific applications. It does not establish responsibility for guest operating systems, database contents or business logic.
For each layer, agree who implements changes, handles incidents and approves decisions. Where several partners are involved, name a contact for the interfaces between them.
Use the following as a starting point for that discussion:
| Area | Responsibility to clarify | Handover evidence |
|---|---|---|
| Infrastructure and platform | Updates, capacity and technical incidents | Documented service scope and contact route |
| Application and database | Releases, configuration and functional errors | Named teams and agreed test procedures |
| Monitoring and alerts | Maintain rules, assess notifications and respond | A test alert reaches the responsible person |
| Backup and recovery | Check backups, restore service and accept the result | Recorded recovery exercise |
| Incident communication | Set priorities and inform affected parties | Tested reporting and escalation process |
The actual allocation depends on the agreed service scope. This table helps expose gaps early; it does not replace a service agreement.
2. Monitor how the service is used
A running VM does not prove that a customer can complete an order. Add checks for important application workflows to infrastructure metrics.
For a customer portal, this might mean checking reachability, using a suitable test account and verifying a defined workflow. Design the permissions and test data for that purpose.
Every urgent alert needs an owner, initial response instructions and an escalation route. A dashboard alone does not trigger a response. Test notification delivery before handover.
Also agree response coverage explicitly. Continuous technical monitoring does not, by itself, establish round-the-clock incident handling.
3. Plan backup and recovery together
A successful backup job initially confirms what the backup process reported. A restore exercise must establish whether those backups can make the application usable again.
Start with two business objectives:
- RPO: What is the maximum acceptable data loss, expressed as a period of time?
- RTO: What is the maximum acceptable time to restore service?
For example, a daily backup does not meet a requirement to lose no more than one hour of data. The recovery procedure must also fit within the available recovery time.
Include the configuration and access needed to recover, alongside business data. Protect backups against failures and access risks that could also affect production.
Test restoration in a suitable, separate environment. Record elapsed time and the recovered data point, check integrity and obtain acceptance of the relevant business workflow. Repeat testing periodically and after significant changes.
4. Maintenance is part of operations
Operating systems, platform components and applications still need maintenance after migration. Agree maintenance windows, approvals and procedures for failed changes.
Capacity also requires ongoing attention: Who detects growing storage demand? Who can request additional resources? Who assesses the cost?
Answer these questions during architecture planning. Otherwise, an environment can work technically while requiring fresh negotiations for every operational change.
5. Accept the handover through practical checks
Agree an observation period after cutover. The migration team, operations and application owners support the environment together. Define its duration, contact coverage and completion criteria beforehand.
Acceptance should make at least the following verifiable:
- Architecture, dependencies and operating instructions are current.
- Responsible people have the required access.
- Alerting, recovery and escalation have been tested.
- Outstanding issues have owners and deadlines.
- The transition from the project team to ongoing operations is explicitly confirmed.
Retire source systems after approval, taking retention needs, remaining dependencies and the agreed rollback plan into account.
Cloud transformation should deliver dependable daily operations
A migration delivers lasting value when the operations team can manage the target environment as conditions change. This requires clear tasks, rehearsed procedures and a documented service scope.
ayedo supports cloud transformation from assessment and architecture through migration to the agreed platform operations. Before operations begin, we clarify which services we provide and how we work with your teams.
Plan your migration and operational handover with us.
More in this series:
- Cloud migration without a big bang
- Proxmox, OpenStack or Kubernetes: Which platform fits which workloads?
Further reading: Google SRE: Monitoring, Define recovery objectives, Plan backups and Test recovery periodically.