Serverless Architecture for 2M Documents a Day

I’ve just published a new article on AWS Builder Center about AWS Lambda. The article explains how Lambda works, how it fits into serverless and event-driven architectures, and some common backend use cases such as APIs, message processing, file processing, and automation. It also covers an important point: serverless does not mean Lambda is the right choice for every workload. 🔗 https://lnkd.in/dB7Nk7Mm #AWS #AWSLambda #Serverless #BackendDevelopment #EventDrivenArchitecture

Digging in the memory lane, found reference architectures on how to deploy RedCap in AWS. 2026 version: https://lnkd.in/ehFPX3WA 2019 version: https://lnkd.in/eRGRKjMt

Moving to AWS doesn't make you cloud-native. After 5 years of working with cloud-native architectures, here's the real shift: ❌ Lift and shift = same problems, now on AWS ✅ Cloud-native = design for failure, scale on demand, decouple everything The mindset change matters more than the services you use. S3, Lambda, SQS, RDS — these are tools. Resilience, observability, and loose coupling — that's the architecture. #AWS #CloudNative #Microservices #SoftwareArchitecture #BackendEngineering

I’ve just published a new article on AWS Builder Center about Amazon SQS. The article explains how message queues help decouple applications, improve resilience, and support asynchronous processing in backend and microservices architectures. I also cover the difference between Standard and FIFO queues, how SQS works with AWS Lambda, and the role of Dead Letter Queues (DLQs) when message processing fails. 🔗 https://lnkd.in/d5rJvw4g #AWS #AmazonSQS #EventDrivenArchitecture #BackendDevelopment #Microservices

Your application can scale—but can your traffic layer keep up? AWS Elastic Load Balancing intelligently distributes incoming traffic across healthy resources, helping applications stay available, responsive, and ready for demand. Build an AWS architecture that scales without creating a single point of failure. Connect with Aishwary Chauhan and Sagar Gupta to design your next scalable workload. #AWS #ElasticLoadBalancing #CloudArchitecture #Scalability #AWSCloud

The 15-minute Lambda barrier is officially history. ⏱️ AWS Lambda now supports up to a 90-minute timeout on Lambda Managed Instances (LMI) for asynchronous and ESM workloads. Long-running batch ETL, AI model evals, and media transcoding no longer need to be pushed into ECS Fargate or AWS Batch just because they exceed 15 minutes. But running serverless compute continuously for 1.5 hours brings a whole new set of failure modes: • 9-hour SQS visibility timeouts • 350-second NAT Gateway idle socket drops • Ephemeral STS credential expiration • The necessity of idempotent design I published a comprehensive guide covering the new execution model, how it works with Durable Functions, and the gotchas you need to avoid. Check out the full article here: https://lnkd.in/dtrRi-Fa help me post this in linkdin #AWS #CloudArchitecture #Serverless #BackendEngineering #AWSCommunity

🚀 AWS Lambda – One of the AWS Services I Keep Coming Back To When I first started working with AWS, Lambda looked very simple: Write some code → upload it → AWS runs it. But as I started understanding real-world cloud architectures, I realized there is much more to Lambda than just running a piece of code. Here are a few things I keep in mind when working with AWS Lambda 👇 🔹 1. Serverless Execution With Lambda, we don't need to manage servers, operating systems, or infrastructure. AWS takes care of the underlying compute. 🔹 2. Event-Driven Architecture Lambda becomes really powerful when it's triggered by events. For example: S3 → Lambda → Process File or API Gateway → Lambda → Application Logic or EventBridge → Lambda → Automation 🔹 3. Automatic Scaling Lambda can scale based on incoming requests, which makes it useful for workloads where traffic can change significantly. 🔹 4. IAM & Least Privilege 🔐 One important practice is giving the Lambda function only the permissions it actually needs. For example, if a function only needs to read from an S3 bucket, I wouldn't give it full S3 access. 🔹 5. Monitoring with CloudWatch I normally look at CloudWatch when troubleshooting Lambda: • Invocation count • Errors • Duration • Throttles • Logs These metrics can quickly help identify where an issue is happening. 🔹 6. Environment Variables & Secrets Configuration can be separated from the code using environment variables, while sensitive information should be managed through appropriate secret-management solutions. 🔹 7. Cold Starts Matter One thing that's important to understand is cold starts. Depending on the runtime, package size, initialization work, and architecture, the first invocation can sometimes take longer. 🔹 8. Keep Functions Small & Focused Instead of putting an entire application into one Lambda function, it's usually better to keep functions focused on a specific responsibility. 💡 My biggest takeaway: Lambda isn't just about "running code without servers." It's about building event-driven, scalable, and automated applications while reducing the amount of infrastructure we need to manage. The more I work with AWS, the more I see Lambda being useful for automation, APIs, file processing, scheduled tasks, integrations, and operational workflows. What was the first real-world use case where you used AWS Lambda? 👇 #AWS #AWSLambda #Serverless #CloudComputing #AWSCloud #DevOps #CloudEngineer #CloudArchitecture #Automation #AWSDevOps #CloudEngineering #Learning #KnowledgeSh

Goodbye 15-minute workaround hacks! 😃 AWS Lambda now supports a 90-minute function timeout for async and event-driven workloads. A boon for those who rely on Serverless Architecture! https://lnkd.in/diEVJKqi

This is a very thoughtful and unique architectural problem, along with an equally interesting solution. Watch all three videos for a clear understanding of the concept and approach.