300 points by cloudengineer 5 months ago flag hide 16 comments
serverlesssam 5 months ago next
Great post! I've been working on a similar project and I was wondering what toolset you used for your serverless microservices? I'm considering AWS Lambda for my setup. Thanks for sharing!
learnbuidserve 5 months ago next
Hey @serverlessSam, I'm glad you enjoyed the post! I actually used the Serverless Framework to manage my AWS resources and deployments. It's made working with AWS Lambda and Kinesis a breeze. I highly recommend it!
nodeprodev 5 months ago prev next
Very clever piece! I've also tried implementing serverless a while back, but had a tough time with cold start issues. Have you encountered similar problems, and if so, how'd you tackle 'em?
learnbuidserve 5 months ago next
@nodeProDev, thanks! Yeah, on certain projects I did face cold start issues in my serverless architecture. To mitigate that, I heavily relied on using provisioned concurrency for my Lambda functions. That helped me out quite a bit, altough it meant adding to our AWS bill
trustytech 5 months ago prev next
Impressive work with the microservices @learnBuidServe! I'm curious, what database tech did you opt for to minimize write latency for your data?
learnbuidserve 5 months ago next
@trustyTech. I chose DynamoDB for the job, as it has a really quick write latency and can have items written to it in the micro-second range. I'm quite delighted with the choice!
codesavvy 5 months ago prev next
Awesome post @learnBuidServe! I'm currently working through a refactoring of my own monolithic app into serverless microservices. Have you contemplated using the 'Strangler' pattern?
learnbuidserve 5 months ago next
Hey, @codeSavvy! I did implement the Strangler pattern in my microservices migration strategy, it worked really well! It allowed me to break down the app into smaller, more manageable parts.
fastfred 5 months ago prev next
Matched set with this scale question: Impressive results @learnBuidServe! I'm handling parity scaling with millions of records on my application. Have you testes the capacity of your solution, and do you have any predictions of the throughput constraints of what you've done here?
learnbuidserve 5 months ago next
@fastFred, many thanks for the feedback! I haven't hit those scaling numbers you mention, but my tests showed I could handle more than 10,000 requests per second. Throughput constraints were highly dependent on my DynamoDB configuration, so a separate study would be needed to answer your question.
serverless50 5 months ago prev next
Well done! Have you looked into other integrations like Azure functions, event grid and logic apps, or simply stuck to AWS?
learnbuidserve 5 months ago next
@serverLess50, it's flattering to hear that! You've asked a great question. I've been using infrastructures across different cloud providers, but for this project, I stuck to AWS purely for its Serverless Application Model (SAM) capabilities, which fit my use case
codefighter 5 months ago prev next
Terrific post @learnBuidServe! I'm preparing to adopt serverless architectures for our production systems. I'm concerned about observability & monitoring in such decentralized approaches. Any tips on how to handle that?
learnbuidserve 5 months ago next
@codeFighter, thank you! I recommend giving AWS X-Ray a go – it helps you analyze and debug distributed applications. It's an excellent tool when it comes to monitoring and observability for serverless.
liftoff50 5 months ago prev next
I'm having troubles scaling with EC2, I may give this serverless approach a whirl. Cheers! @learnBuidServe
learnbuidserve 5 months ago next
@liftOff50, I'm glad to hear it piqued your interest! I had my fair share of struggles with scaling on EC2. Good luck with your serverless journey!