Introduction

Schema is structure/definition of data, defines:

• What fields exist (id, name, email)
• Their types (int, string)
• Rules (required? optional?)

Example:

{
  "id": 1,
  "name": "Min Hen",
  "email": "[email protected]"
}

In microservices, services usually communicate through events (Kafka):

• Multiple services consume the same data
• Data evolves over time
• It is hard to maintain each schema in each service to fit latest schema one by one

Task Service sends:              Orchestrator expects:
{                                {
  "task**Id**": "123",                 "task**_id**": "123",
  "taskName": "Process",           "name": "Process",
  "type": "COMPUTE"                "taskType": "COMPUTE"
}                                }

Issues:

• Field mismatch (taskId vs task_id)
• No compile-time validation
• Runtime failures (null values)
• Silent bugs (hard to debug)