Integration Boundaries, Events, and the Outbox Pattern Enterprise systems integrate with banks, biometrics, email/SMS, government services, accounting systems, suppliers, and identity providers. External side effects cannot participate in the same database transaction.
Business - - - - Inbox/idempoten Outbox row Publisher/worker External system transaction > > > > cy
Boundary rule • Do not call an external provider halfway through a database transaction and assume both systems commit atomically.
• Persist the business change and an outbox/integration intent atomically; a worker publishes/sends after commit and records delivery state.
• Inbound webhooks should verify authenticity, store a stable provider/event ID, and process idempotently so duplicate delivery cannot duplicate business effects.
Event design • Use business events such as PurchaseOrderApproved or PayrollRunFinalized, not low-level "row updated" messages as the default integration contract.
• Keep event payloads versionable and avoid leaking internal ORM row shapes to external consumers. BEGIN TRANSACTION update purchase_order set status = 'APPROVED' insert audit_event (...) insert integration_outbox (event_type, aggregate_id, payload) COMMIT
worker -> supplier/accounting integration
The outbox solves "database committed but message was never sent" by making publish intent durable. Source basis: transactional outbox/inbox architectural pattern; external systems cannot share a local PostgreSQL atomic transaction.
Enterprise systems are difficult because business invariants, approvals, auditability, effective dates, and permissions interact over long-lived data.
Restate the goal of “Integration Boundaries Events and Outbox” as one observable behavior. Write the success signal before touching code.
Trace the boundary map: business workflow → invariants → permissions → transaction/audit → operational outcome. Mark the one boundary this lesson changes.
Inspect the current repository state and record the exact files, route, table, or runtime involved. Avoid creating an abstraction before you can name the existing owner.
Implement the smallest vertical slice. Add one coherent change, run the narrowest verification command, and read the output before continuing.
Connect the slice to TaskFlow’s phase artifact: Enterprise expansion blueprint for an HRIS/procurement/inventory/payroll-style system with modules, workflows, events, audit, and scaling triggers.. Verify behavior from the outside as a user or API client would.
Create one controlled failure that is safe to reproduce. Predict the symptom first, capture evidence, locate the earliest broken boundary, then fix only that cause.
Close the reference and rebuild or explain the core change from memory. Record what you would monitor or test before trusting the same change in production.
Requisition -> Approval -> Purchase Order -> Goods Receipt -> Invoice
Invariant examples:
- received quantity <= ordered quantity
- finalized approvals are auditable
- closed orders reject new receiptsDo not copy this as a finished answer. Predict what each boundary does, type the smallest useful slice, run it, then compare the result with your prediction.
Return to the nearest observable boundary: input, network, process log, database, or rendered UI. Compare actual evidence with the predicted path.
Change one variable at a time and reproduce the original failure after the fix; otherwise you have correlation, not diagnosis.
Without copying the final source, create a second small scenario that applies “Integration Boundaries Events and Outbox” to a different TaskFlow resource or input. Write the acceptance criteria first, implement it, and provide the evidence that each criterion is satisfied.
Design modules with explicit ownership and public interfaces inside a modular monolith.
Use the smallest change that demonstrates this lesson. Stop after it works and capture evidence.Shared database tables with unclear ownership across domains
Predict the symptom before causing the failure. Then locate the earliest boundary where reality diverges from expectation.Why does this responsibility live here, what input does it trust, what output proves success, and what would go wrong if another layer owned it?
If your explanation is vague, repeat the boundary trace before moving on.Close the reference. Reproduce the core idea from memory or explain the exact request/data flow aloud. Then check your diff or notes for accidental shortcuts that bypass validation, authorization, error handling, or verification earned in earlier phases.