Vadym Kazulkin 🇺🇦’s Post

I benchmarked Quarkus (JVM + Native) vs .NET 10 on AWS Lambda with real PostgreSQL connections. Same VPC, same memory, same query. Quarkus Native delivers 472ms total billed cold start (3x less than .NET, 5x less than JVM), 89ms first execution, and just 67MB of memory. At scale, that translates directly into lower AWS bills and better response times. #AWS #Lambda #Serverless #Java #Quarkus #DotNet #ColdStart #CostOptimization

Day 60 of building in public 🚀 Today I deployed my Scalable URL Redirect Service built with Spring Boot + PostgreSQL on an AWS EC2 server. What the system does: • Accepts a long URL through a REST API • Generates a unique short code • Stores mapping in PostgreSQL • Redirects users instantly when the short URL is accessed Tech stack used: Java • Spring Boot • PostgreSQL • REST APIs • AWS EC2 • Gradle Key things I learned while building and deploying this: • Designing clean REST endpoints • Working with JPA repositories and database persistence • Handling HTTP redirects in Spring Boot • Deploying backend services on a cloud server • Debugging real deployment issues like port conflicts and DB connections Seeing the API running live on a server and successfully redirecting URLs was a great learning moment. Next steps: • Add Redis caching • Improve short code generation strategy • Implement rate limiting #Java #SpringBoot #BackendDevelopment #AWS #PostgreSQL #BuildInPublic #SoftwareEngineering

5 years. Here's what I've shipped: → 20.5B+ events/year processed with millisecond   SLA across 8 global marketplaces → 40% reduction in API P95 latency through async   microservices and Redis caching → 99.99% uptime on AWS-hosted distributed systems → CI/CD cycles cut from weeks to days → 90%+ test coverage maintained across all projects Built on: Java · Spring Boot · Kafka · AWS ·  Kubernetes · Terraform Always open to conversations about distributed  systems and cloud-native engineering. #Java #DistributedSystems #AWS #SpringBoot #Kafka

Java on AWS: The Cold Start Myth “Java is slow on AWS” is an outdated argument. Most cold-start issues come from: - Fat JARs - Excessive dependencies - Heavy static initialization Modern JVM + good architecture: - Starts fast - Scales well - Costs less than people expect Java isn’t the problem. Architecture is. 💡 Takeaway: Cloud cost is an engineering problem, not a language problem. #Java #AWS #CloudComputing #BackendEngineering

🚀 Built a Full AWS-Integrated Backend with Spring Boot! After a deep debugging journey, I successfully connected: 🔹 Spring Boot application 🔹 Amazon S3 (for file storage) 🔹 Amazon DynamoDB (for structured data) 💡 Key learnings from this experience: ✅ Understanding AWS credential flow (IAM, Access Keys, Profiles) ✅ Fixing real-world errors like: BeanCreationException CredentialsProvider issues ResourceNotFoundException (missing tables) ✅ Designing clean configuration using profiles (local vs dev) ✅ Integrating S3 + DynamoDB in a scalable way ⚙️ Tech Stack: Spring Boot | AWS SDK v2 | Amazon S3 | Amazon DynamoDB | Maven This wasn’t just coding — it was about debugging, understanding architecture, and thinking like a backend engineer. #SpringBoot #AWS #BackendDevelopment #DynamoDB #S3 #Java #CloudComputing #LearningByDoing You can find this code on : https://lnkd.in/gXmRZWAZ

In my first post, I presented an analysis and a solution for a case of transferring large volumes of files using Java Spring and AWS cloud services. https://lnkd.in/d7JXzMgR

Check out the example of AWS Lambda with managed Java 25 runtime using Spring Boot 4 with AWS Serverless Java Container, Hibernate ORM framework, Hikari Connection Pool, and Amazon Aurora DSQL database. 👉 https://lnkd.in/e-2WYWMf If you like my content, please follow me on GitHub (github.com/Vadym79) and give my repositories a star! Please also check out my website 👉 vkazulkin.com for more technical content and upcoming public speaking activities. Amazon Web Services (AWS) #Java #Serverless #Spring #SpringBoot #PostgreSQL #AuroraDSQL

Serverless with AWS Lambda in 60 seconds ⚡ Build and run applications without worrying about servers. With AWS Lambda, you simply upload your code while AWS automatically manages the infrastructure, scaling, and uptime. ⚙️ Execute code in response to events from Amazon S3, Amazon API Gateway, or scheduled triggers 💻 Works with Python, Node.js, Java, Go, and several other languages 💰 You’re charged only for the time your code runs Whether it’s processing files, automating workflows, or powering APIs, Lambda enables a flexible and efficient serverless approach. 💬 How are you using Lambda in your projects? Let’s discuss below! #AWS #Lambda #Serverless #CloudComputing #TechEducation #DeveloperLife #Coding #DevOps #CloudNative #WebDevelopment #certificate #AWSTraining AWS Training & Certification

𝗘𝗖𝟮 𝘃𝘀 𝗘𝗖𝗦 𝘃𝘀 𝗟𝗮𝗺𝗯𝗱𝗮. Every Java developer gets asked this. Most give a vague answer. Here is how to think about it simply: 🏠 EC2 — You rent a house. You manage everything. 🏢 ECS — You rent a flat. AWS manages the building. 🏨 Lambda — You book a hotel room. Pay only when you use it. Swipe for the full breakdown — with the exact answer your interviewer wants to hear. 👇 Save this before your next system design round. 🔖 #AWS #Java #SpringBoot #SystemDesign #CloudComputing #DebugWithPurpose

AWS Lambda now tells you which Availability Zone your function is running in. Before this update, Lambda would run across multiple AZs automatically but your function had no way to know which one it was in. This caused silent cross zone calls to services like ElastiCache and RDS which added latency and unnecessary cost. Now AWS has added a metadata endpoint inside every Lambda execution environment. Lambda automatically sets two environment variables in every execution environment. You do not configure anything manually. AWS_LAMBDA_METADATA_API gives you the server address and AWS_LAMBDA_METADATA_TOKEN gives you a unique auth token for that execution environment. The actual endpoint to call is "http://AWS_LAMBDA_METADATA_API/2026-01-15/metadata/execution-environment" You pass the token as Authorization Bearer header in the request. This token based auth. The response gives you back the AZ ID like use1-az1. The recommended way to use this is through Powertools for AWS Lambda. It supports Python, TypeScript, Java and .NET. Powertools handles response caching automatically after the first call and also handles SnapStart cache invalidation when your function restores in a different AZ. You do not have to manage any of that yourself. Once you have the AZ ID your function can route calls to same zone endpoints of ElastiCache or RDS and avoid cross zone traffic completely. You can also use it for AZ specific fault injection testing. Works with all runtimes including custom runtimes and container images. Compatible with SnapStart and Provisioned Concurrency. Works whether your function is VPC enabled or not. No additional cost. Read more here https://lnkd.in/gViFDvuY Documentation https://lnkd.in/gM9y4M_4 #AWS #LAMBDA