Skip to main content

Handover

Handover is complete when the customer's team can operate, change and recover the system without us — and can prove it. A document transfer is not a handover.

Preparation​

AreaRequirement
DeliverablesInfrastructure, applications, backups and monitoring deployed and working
DocumentationDeployment, operations, troubleshooting and recovery procedures complete
TestingFinal functional, load and recovery tests passed
AccessCustomer has access to every account, repository, pipeline and state store
KnowledgeTraining sessions prepared and scheduled
RunbooksWritten, reviewed, and linked from the alarms that trigger them

Handover documentation​

Project overview​

  • Scope and what was delivered.
  • Architecture diagram with each component labelled.
  • Environments and how they differ.
  • Key services and why each was selected.

Infrastructure as code​

ContentDetail
RepositoriesLocation, purpose, ownership
Module structureWhat each module provisions
StateBackend configuration, bucket, lock table, access
EnvironmentsVariable files and how to apply them
AccessWho can apply, and how approval works
Change processHow a change is proposed, reviewed and applied

Application delivery​

  • Pipeline description, stage by stage.
  • How to build and deploy manually if the pipeline is unavailable.
  • How to roll back, and how long it takes.
  • Where artefacts are stored and how they are tagged.
  • Environment variables and secrets, and where they are held.

Operations​

  • Dashboards and what each shows.
  • Alarms and the runbook for each.
  • Routine tasks: certificate renewal, key rotation, patching, scaling.
  • Backup schedule and restore procedure.
  • Incident process, escalation and contacts.

Security​

  • IAM model and how access is granted.
  • Network segmentation.
  • Encryption and key management.
  • Detection configuration.
  • How to respond to a security finding.

Recovery​

  • RTO and RPO agreed.
  • Backup locations and retention.
  • Restore procedure, with measured times from the drill.
  • Failover procedure.
  • What to do if the primary region is unavailable.

Knowledge transfer​

SessionAudienceContent
Architecture walkthroughEngineering and operationsComponents, data flow, decisions and trade-offs
Pipeline walkthroughEngineeringBuild, test, deploy, rollback; how to change it
Operations walkthroughOn-callDashboards, alarms, runbooks, routine tasks
Security walkthroughSecurity and operationsControls, detection, response
Recovery drillOperationsExecute a restore and a failover, measured
Change processAllHow change is proposed, reviewed and approved

Sessions are recorded and the recordings handed over, so the customer can train new joiners.

Verification​

Handover is verified, not assumed:

  • The customer's engineer deploys a change through the pipeline unaided.
  • The customer's engineer rolls it back unaided.
  • The customer's engineer restores a backup into a test environment.
  • The on-call engineer responds to a simulated alarm using the runbook.
  • The customer can access every repository, account and state store.
  • Every alarm resolves to a runbook.
  • Documentation has been reviewed by the customer and gaps corrected.

A handover where the customer watched but did not do has not transferred capability.

Acceptance​

ItemAccepted byEvidence
DocumentationTechnical ownerReview record
TrainingAttendeesAttendance and recordings
Operational readinessOperations leadDrill results
Security readinessSecurity leadControl verification
Formal sign-offSponsorSigned acceptance

Post-handover​

PeriodSupport
Hypercare (2–4 weeks)Daily check-in, priority response, defect fixes
Transition (1–3 months)Named contact, advisory support, escalation path
Steady statePer the support agreement, if any

Hypercare ends on a defined date with a defined exit criterion — for example, two weeks without a critical incident and a completed drill.