Documentation

CAP Operator takes the complexity out of running SAP Cloud Application Programming Model (CAP) applications on Kubernetes — so you ship features, not infrastructure.

Whether you’re building a full SaaS platform with hundreds of tenants or a lean shared service with no tenant lifecycle at all, CAP Operator handles the heavy lifting:

• Multi-tenant applications — end-to-end SaaS lifecycle: onboarding, upgrades, deprovisioning, and everything in between, wired directly to SAP SaaS Provisioning service.
• Services-only (tenant-agnostic) applications — expose workloads publicly via named subdomains with zero tenant subscription overhead. Perfect for shared APIs and platform services.

Deploy fast, deploy clean

• One-command deployment of CAP servers, Approuters, and supporting workloads — networking included.
• Content deployers run as ordered Kubernetes jobs on every release, so your HTML5 apps and service configurations are always in sync.
• Services-only mode lets you go live without an Approuter or subscription flow, just clean subdomain-based exposure.

Own your tenant lifecycle

• Asynchronous tenant onboarding and offboarding, driven by SAP SaaS Provisioning service and executed as Kubernetes jobs.
• Automated tenant upgrades the moment a new version is ready — no manual rollouts.
• Fully sequenced operation pipelines with pre/post hooks and per-step failure control, so complex provisioning logic stays readable and reliable.
• Subscription dependency callbacks handled automatically, with per-service opt-in control.
• Pin individual tenants to a version when you need to — upgrade on your terms.

Networking that just works

• Automatic TLS certificate provisioning and DNS entry management, including customer-specific domains.
• Namespace-scoped and cluster-scoped domain resources — share a domain across applications or keep it isolated.
• Mutual TLS and Go-template-based custom DNS generation for advanced networking scenarios.

Built for production operations

• Credential rotation triggers automatic rolling restarts of only the affected workloads — everything else keeps running.
• Prometheus-driven cleanup automatically retires unused application versions, keeping your cluster lean.
• Version affinity (experimental): users stay on the version they started with until their session ends, making zero-downtime upgrades a reality.
• Rich Prometheus metrics across the controller and subscription server — reconciliation errors, tenant operation durations, subscription request rates, and more.

The following diagram shows the major automation steps handled by CAP Operator during application deployment:

Explore the following sections to learn more.