From Monolith to Kubernetes Workloads
We finally made the leap: moved a legacy Java monolith to Kubernetes. Here’s what helped: Broke out background jobs as independent CronJobs Converted WAR to fat JARs using Spring Boot Used ...
We finally made the leap: moved a legacy Java monolith to Kubernetes. Here’s what helped: Broke out background jobs as independent CronJobs Converted WAR to fat JARs using Spring Boot Used ...
As GraphQL picked up adoption, schema versioning surfaced as a practical challenge. REST gave us /v1/, /v2/ endpoints. But with GraphQL, you only get one endpoint — so how do you evolve a schema w...
I used to fear date and time handling in Java — until the java.time API landed in Java 8. This quarter, I finally explored the TemporalAdjusters and custom parsing capabilities. Here’s a quick exa...
While JSON is human-readable and flexible, it’s often not the best choice for internal service-to-service communication. Why Protobuf? Smaller payload sizes Faster (de)serialization Strong...
Safari — the final boss of frontend compatibility. We faced issues with: Date pickers not firing onChange CORS preflight caching behaving inconsistently Font rendering glitches with SVG W...
Spring Boot 2.0 hit GA and came with subtle but powerful refinements: Embedded Netty support for WebFlux Actuator improvements (more endpoints + security config) Java 8+ baseline I found t...
JSON is great for human readability, but once you’re exchanging high volumes of messages between internal services, its verbosity becomes a real issue. That’s where Protocol Buffers (Protobuf) com...
Legacy systems aren’t inherently bad — they’re often successful systems that adapted over time. But modernizing them is rarely straightforward. The Strangler Pattern offers a gradual path to repla...
Tech Radar this quarter emphasized: Protobuf replacing JSON in backend systems Chaos testing tools like Chaos Monkey gaining mainstream Move toward declarative security (OPA, etc.) The li...
Couldn’t attend in person, but followed most keynotes and sessions online. Highlights: Amazon SageMaker democratizing ML pipelines Aurora Multi-Master and Serverless variants Fargate elimin...