Managed Services
Service Quotas
Summary
A managed service's relevant quotas are identified before production use and monitored so capacity or demand can be adjusted before a quota is reached.
Standards
- A managed service's quota relevant to a workload's expected demand MUST be identified before the workload depends on that service in production.
std-plat-service-quotas-01 - A managed service's quota usage MUST be monitored against expected demand so capacity or workload can be adjusted before the quota is reached.
std-plat-service-quotas-02
Implements These Principles
Availability
Summary
A managed service uses redundancy across failure domains according to the workload's defined RTO.
Standards
- Where a workload's defined RTO cannot tolerate the loss of a single failure domain, a managed service it depends on MUST have its redundancy across failure domains enabled.
std-plat-availability-01 - A workload whose defined RTO can tolerate the loss of a single failure domain SHOULD use a managed service's lower-redundancy configuration.
std-plat-availability-02
Implements These Principles
Backup & Recovery
Summary
A managed service's built-in backup meets the workload's RPO and RTO, and its restoration is verified during disaster recovery testing.
Standards
- Where a managed service's built-in backup is relied on instead of a separate one, its frequency and retention MUST meet the workload's defined RPO and RTO.
std-plat-backup-recovery-01 - Restoration from a managed service's built-in backup MUST be verified as part of disaster recovery testing.
std-plat-backup-recovery-02
Related Standards
Implements These Principles
Automatic Patching
Summary
Automatic patching is normally enabled, with the maintenance window set to a period of low impact to users.
Standards
- Automatic patching SHOULD remain enabled for a managed service instance.
std-plat-automatic-patching-01 - A managed service instance's maintenance window SHOULD be set to a period of low impact to end users.
std-plat-automatic-patching-02
Implements These Principles
Observability Integration
Summary
Telemetry and alerts required to operate a managed service are integrated with the organisation's centralised observability and alerting platforms.
Standards
- Telemetry required to operate a managed service, including applicable logs and metrics, MUST be integrated into the organisation's centralised observability platform.
std-plat-observability-integration-01 - A managed service's own native alerting SHOULD be configured to notify through the organisation's centralised alerting mechanism.
std-plat-observability-integration-02