Each row scores how tightly two systems are coupled on one dimension, from data and timing through to platform and transaction, on a shared one-to-five scale. The lower the bars, the looser and more resilient the integration.
Tight coupling on any row is where a change in one system forces a change in the other. The tallest bar is the dimension to loosen first to make the integration easier to evolve safely.
One is loose, five is tight. The radar redraws with each score.
Use a topic-based pub/sub broker (Kafka, SNS/SQS, Azure Service Bus) so the source publishes domain events and the target subscribes independently. Add a dead-letter queue for failed messages.
The coupling profile is mixed. Pub/sub decouples timing and platform while keeping data contracts explicit through event schemas.
