Skip to main content

Change management

Every production change to Sally goes through the same tracked path: recorded, risk-assessed, tested, approved by a second pair of eyes, deployed in a controlled window, verified after go-live, and reversible. The programme aligns with ISO 27001 control A.8.32 (Change Management).

Ticketing
Azure DevOps
Version control
Git
Regular rollout
Weekly, Saturdays
Approval
Four-eyes principle

What we cover as a "change"​

Backend and frontend code, infrastructure and cloud resources, databases and storage, security and configuration components, integrations and third-party interfaces, deployments and releases, and changes to operational or support processes with technical impact.

The seven principles​

✓
Traceability
Every production-relevant change is documented before implementation, versioned, and logged after implementation.
✓
Risk-based approach
Impact on security, data protection, availability and functionality is assessed before the change proceeds.
✓
Test before production
Changes are exercised before they are released into production.
✓
Environment separation
Development, test and production environments are kept apart.
✓
Four-eyes principle
Production changes require sign-off by technical leadership under four-eyes.
✓
Rollback capability
Production changes must have a technical path back to a suitable prior state.
✓
Impact minimisation
Changes are planned and executed to keep operational disruption as low as possible.

The release cadence​

  • Weekly rollout every Saturday. The full system is rolled out on a fixed weekly rhythm.
  • Mid-week fixes and small updates. Bug fixes and minor adjustments can additionally deploy during the week.
  • Off-hours deployment. Mid-week deployments run in the evening or at night to minimise disruption.
  • Critical exceptions. In critical exceptional cases, deployments can also happen outside these windows.

Pre-go-live testing​

Every production change goes through the same checklist:

  • Two manual functional tests. One by the developer, one by an independent tester.
  • Automated end-to-end tests. Core user flows are covered: account creation, login, onboarding, calendar integration, account deletion, and further central process paths.
  • Domain- and technical review. A targeted review of the change in the context of the affected components.

Deployment and verification​

Deployments run through Azure DevOps. After go-live we check that the affected functions work, that no obvious follow-on errors surface, that the platform stays stable and reachable, and that no security or data protection side-effects are visible.

Emergency changes​

Security-critical fixes, severe defects and acute production incidents follow an accelerated path. The principle of controlled execution still applies: documentation, follow-up review and approval confirmation happen immediately after the emergency change.