PATTERN Cited by 1 source
Use-case-based provisioning¶
An infrastructure provisioning pattern where engineers specify what they're trying to accomplish (in plain language or from a use-case catalog) rather than which resources they need. The platform maps intent to a configured infrastructure template.
Structure¶
- Input: natural-language description or catalog selection of purpose (e.g., "demo for a financial services customer", "reproduce support issue", "hackathon session")
- Mapping layer: matches intent to a pre-configured template from a governed catalog
- Output: fully-configured environment tailored to the stated purpose
Rationale¶
Resource-based provisioning ("give me a workspace with X vCPUs in us-east-1") pushes configuration burden onto engineers who may not know the optimal setup for their use case. Use-case-based provisioning inverts this: the platform owns configuration knowledge, engineers own intent.
Benefits¶
- Engineers don't need to know infrastructure details
- Consistent configuration per use case (every "financial services demo" gets the same governed setup)
- Natural integration with AI agents — agents describe intent in language, not in IaC syntax
- Template updates propagate to all future provisioning without user changes
Trade-offs¶
- Template catalog must be maintained and expanded as use cases evolve
- Edge cases that don't fit existing templates require platform-team intervention
- Abstraction can hide important configuration when engineers do need control
Seen In¶
- sources/2026-07-23-databricks-self-serve-infrastructure-vending-machine — "You describe what you're trying to do… and you get an environment configured for that purpose"; core design principle of Databricks FEVM