AI Prompt for Microservice Extraction

Extract microservices safely with this AI prompt: service boundaries, API contracts, data migration strategy, and reversible strangler steps included.

The prompt

Role: You are a systems architect who extracts microservices from monoliths without big-bang rewrites.

Context: Monolith: [STACK]. Code to extract: [CODE_OR_MODULE]. Intended service responsibility: [SERVICE_RESPONSIBILITY]. Current coupling: [COUPLING: shared DB / shared code / none].

Task: 1) Define the service boundary: what goes in, what stays out, and the API contract (endpoints/messages with schemas). 2) Design the data strategy: shared DB initially vs own DB with migration plan — justify the choice. 3) Write the extraction plan as strangler steps: each step independently deployable and reversible. 4) List the new failure modes (network, partial failure, eventual consistency) and how each is handled. 5) Define when the extraction is DONE (old code deleted, traffic fully shifted).

Format: Boundary + contract first, then data strategy, then the step plan, then failure modes, then done-criteria.

Constraints: No big-bang cutover — every step must be reversible. Be explicit about consistency trade-offs; distributed transactions need justification. The API contract comes before code.

Variables to fill in

Replace each [variable] in the prompt with your own details:

[STACK] [CODE_OR_MODULE] [SERVICE_RESPONSIBILITY] [COUPLING]

Example input

CODE=monolithic order processing function (validate → charge → inventory → notify), TARGET=standalone microservice, STACK=Node.js, BOUNDARIES=keep DB shared initially

How to use this prompt

Describe the current boundaries (shared DB? shared code?) — they constrain the extraction. Define the service's responsibility crisply; vague boundaries create distributed monoliths.