Secure Software Deployment, Operations, Maintenance
From build artifact to steady state
This domain covers getting software into production safely and keeping it healthy. Secure deployment means hardened baselines, configuration management and infrastructure as code under version control, separation of development, test, and production with no production data in lower environments, and controlled release through change management with approval, rollback, and audit trail. Secure the pipeline itself: least-privilege build credentials, secrets in a vault, signed artifacts, and verification that what was tested is what ships.
Operations means monitoring and logging that feed detection: centralized logs, synchronized time, alerting, and a SIEM. Know the incident response lifecycle - preparation, detection and analysis, containment, eradication, recovery, lessons learned - and that evidence handling matters.
Maintenance means vulnerability and patch management on a defined cadence, with emergency patching that still travels a documented change path. Backups must be tested for restore; recovery plans state RTO (time until service is restored) and RPO (tolerable data loss). Retire systems deliberately: revoke credentials, archive data, and destroy media per policy.
Traps
- Logging is not monitoring; collection without alerting detects nothing.
- Patching does not bypass change control; it uses an emergency path with documentation.
- RTO and RPO are routinely swapped in distractors.
- Decommissioning is a security event: orphaned accounts and forgotten data stores are common findings.
Rule of thumb
Before go-live ask three questions: can we detect an attack, can we restore within RTO, and can we patch within our SLA? Any 'no' means the deployment is not ready.