Onlihub
Technical ownership of a commerce platform
A store connection should lead to a usable business process: products match, inventory is understandable, and orders can move toward fulfillment. As CTO at Onlihub, I work across the domain model, backend and integration boundaries. The difficult part is keeping several independently changing systems consistent enough to operate.
Model the product-to-listing and order-to-fulfillment relationships. Keep external IDs and provider capabilities explicit. Decide which side owns inventory, price and each order status before implementing synchronization.
Record intent, assign bounded worker ownership and preserve retry state. The Onlihub case shows idempotency keys, an outbox and PostgreSQL job leases. A local job claim does not guarantee exactly-once effects at an external provider.
Webhooks provide timely signals; periodic checks and explicit failure states help reconcile missed changes. Separate provider adapters from domain services so marketplace-specific behavior does not spread through the whole application.
Real-time everywhere is rarely the right promise. Agree on acceptable delay, API limits and recovery rules per operation. Public integration listings in the linked cases verify the products; they are not personal certifications.
Technical ownership of a commerce platform
A shared model across independent commerce channels
Different data paths for different workloads
Making commerce activity visible on a map
Name the platforms you need to connect, the data that must move and where synchronization currently breaks. Include rough volumes if available; credentials are not needed to discuss the architecture.
Discuss the task