Overview
A service level is more than a percentage. The measured service, time window, exclusions and handling process need to be defined together.
This is a service-level planning guide, not a standalone SLA or a new compensation commitment.
Define the measured service
Separate responsibilities for DNS, edge delivery, origin connectivity and the application. Request failures may arise at different stages and should not automatically be assigned to one service.
Agree on measurement and evidence
Define observation locations, periods, success criteria, data sources and maintenance treatment. Retain timestamps and request evidence rather than deriving monthly availability from one screenshot.
Agree on incident communication and recovery
Define ownership for reporting, diagnosis, origin changes and recovery validation. Prepare fallback, rollback and notification paths in advance and test them within the actual service scope.
Use the actual agreement for remedies and limits
Availability targets, exclusions, claim windows, remedies and limits belong in the agreed documents. This page does not establish a universal availability guarantee or a fixed compensation schedule.