Microservices with Spring Boot: What I Learned at Audi
Practical lessons from several years of backend development in a large, regulated automotive environment.
Context
Since October 2021, I've been working on backend microservices for Audi on the AWS cloud – Java 11, Spring Boot, GraphQL via Node.js/Apollo, and a consistent focus on quality.
1. Resilience is not a feature, but an attitude
Resilience4J, Circuit Breaker, Timeouts, and Bulkheads are not optional extras. They are part of the specification. Anyone who builds them in "later" is accumulating debt that becomes expensive.
2. Observability first
Kibana, Splunk, Prometheus, and NewRelic – no matter which stack: without structured logs, meaningful metrics, and trace IDs, debugging is guesswork.
3. CI/CD is the real architecture
Jenkins, JFrog, Bitbucket Pipelines, and OWASP dependency-check ensure that every change reaches production safely and quickly. The pipeline is part of the product.
Conclusion
Good microservices don't emerge from frameworks, but through discipline and clear responsibility.