Keep the environment controlled after go-live.

Response time and resolution time are different measurements, and an agreement should define both. Response is the interval before a reported fault is acknowledged. Resolution is the interval before normal service returns, and a documented workaround can satisfy it. Restoring service and correcting the cause are separate records under ITIL: an incident is an unplanned interruption to a service, while the underlying cause is recorded as a problem. Any correction that reaches live systems is handled as a change. Priority labels such as P1 and P2 are defined by each agreement rather than fixed by any standard, so the definition behind the label matters more than the number.

A maintenance agreement is only as clear as its schedule of covered assets and its list of exclusions. The same document should state service windows and the customer dependencies the supplier relies on. It should also say how change is approved, since work that alters a covered asset can move it outside the agreed scope. Two figures should be set by the organisation before any of this is priced: the recovery time objective, the longest acceptable outage, and the recovery point objective, the largest acceptable data loss measured in time.