A definition of acceptable levels of downtime has to be agreed between all interested parties. Whilst the ideal goal is a consistent 24/7/365 service provision, this is not always a realistic target due to the potential need for disaster recovery, maintenance or upgrades. As well as agreeing on the acceptable levels of (non)availability, contingency plans need also to be agreed in order that operations can continue seamlessly and safely when any downtime is required.
Managing the availability is directly linked to two other non-functional aspects which are to be assessed to determine the effectiveness of the system:
· Recovery / recoverability (e.g. mean time to recovery - MTTR)
· Reliability (e.g. mean time between failures - MTBF)
Monitoring and measuring these provides a means to identify where improvements are required in terms of stability or quality.