Real-time middleware: three systems, one stream of data
Measured outcomes
Reported outcomes and the attribution attached to this project record.
AttributionThis middleware was built by Kama Adhitia, our core team's ERP Engineering Lead, at a previous employer before Invero was founded.
- 01processing time saved
- > 40%
- 02recurring incidents after observability went live
- -40%
- 03systems connected in real time: ERP, KYC, marketplace
- 3
Context
Business processes ran across three separate systems: an ERP, a KYC verification service, and a marketplace. Each held its own data, and moving information between them had become the slowest point in processing.
Challenge
Data had to move between the three systems in real time without losing consistency. The integration also required monitoring, tracing, and error records for incident diagnosis.
Solution
Before Invero was founded, Kama Adhitia built real-time middleware in Golang and FastAPI. Data moved through Kafka and RabbitMQ, backed by Redis. Observability with SigNoz and Sentry reduced recurring incidents by 40%, while cross-system processing time fell by more than 40%.
Architecture & stack
Technologies and system components named in this project record.
- 01GolangComponent
- 02FastAPIComponent
- 03KafkaComponent
- 04RabbitMQComponent
- 05RedisComponent
- 06SigNozComponent
- 07SentryComponent
Attribution
This middleware was built by Kama Adhitia, our core team's ERP Engineering Lead, at a previous employer before Invero was founded.