Understand the availability commitment
Identify the component being measured, the measurement period and method, planned maintenance treatment and exclusions. Facility, network, virtual machine and application availability describe different boundaries.
Confirm any remedy, reporting process and conditions in the actual agreement. A marketing description or an upstream provider’s commitment should not be assumed to apply unchanged to your complete application.
Define the support route and coverage
Record the support channels, operating coverage, severity definitions and escalation steps. Distinguish acknowledgement, initial response, restoration and final resolution. They are different milestones.
A website enquiry receipt acknowledges a submitted request. It does not establish an incident priority, replace a contracted emergency channel or promise a fixed response time.
Make dependencies and customer duties visible
Specify which monitoring, patching, backups and applications are covered, and which tasks remain with the customer or another provider. Clarify access and approval requirements and how an unresponsive dependency affects the work.
Maintenance, capacity limits, supplier incidents and configuration changes can affect the service. Review relevant notification, change and escalation arrangements before ordering.
Keep recovery objectives separate
Availability design and recovery planning address different failure scenarios. Agree recovery point and recovery time objectives against business needs, then confirm whether the service design and testing support those objectives.
Avoid treating a target as a demonstrated capability without evidence. Define the restoration sequence, dependencies, test frequency, acceptance checks and responsibility for recording the result.
