“We handle infrastructure” is too broad if nobody knows who patches a server, approves a reboot, reviews a failed backup, or follows up with a software vendor. Name the owner, approver, escalation route, and evidence of completion.
Include how each party sees tickets and documentation, and how access is removed when the arrangement ends. Any availability outside normal service hours needs a separate, explicit agreement.
- Ticket intake and routing
- Change approval and maintenance windows
- Alert review and recovery responsibilities
- Shared documentation and access control