Data contracts: schemas and bad data
Kafka moves bytes; it never checks what they mean. This module makes the meaning explicit: an Avro schema for every meter reading, a registry that stores the schemas and refuses breaking changes, and a plan for the records that slip through anyway.
By the end of this module you can: - write and read Avro records with the Schema Registry, and read a record's wire format; - explain subjects, schema ids, and how schemas should be registered in production; - choose a compatibility mode, make safe schema changes, and deploy them in the right order; - recognise a poison pill, park bad records in a dead-letter topic with useful headers, and replay them; - tell bad data from temporary failures, and retry the second kind instead of parking it.
Lessons
- Schemas and the Schema Registry
- Evolving a schema without breaking readers
- Poison pills and dead-letter topics
Then practise with the exercises and check yourself with the module quiz.