See the original LinkedIn post here.
I'd like to share a project I've been working on in past couple of years. In simple words: it's an implementation of shared database transactions for architectures based on microservices.
It can be used to try and solve problem scenarios such as the ones below:
- Spring Boot + JPA => Begin Tx via @Transactional => Read/Write locally, Read/Write API in same Tx => Commit/Rollback Tx.
- Begin Tx via gRPC => Read API1, Read API2, Write API1, Read API1, Read API2 => Commit/Rollback Tx via gRPC.
Other than those scenarios above, it might be helpful to:
- Modernize monolith systems by function while keeping existing centralized RDBMS.
- Share DB transactions on pure microservice/function-based architectures.
- Store weights from dense neural networks into external RDBMS.
- Run parallel hyperparameter tuning epics in different transactions, given (3) is feasible.
- Share DB transactions on AI code agent systems.
Have a good day!
true \
&& ./mvnw clean installDon't forget to change your version number.
true \
&& source .env \
&& ./publish.sh -v "0.0.0-0-SNAPSHOT" \
-a $GCP_SVC_ACCOUNT \
-p $GCP_PROJECT_ID \
-r $GCP_REGION \
-m $MAVEN_REPOSITORY \
-d $DOCKER_REPOSITORY