edit: forgot to mention StackQL [3], which is somewhat similar but takes a SQL-first vs python-first approach. The IaC space is about to get a lot more interesting.
The "dev mode" where you can change lambda code live is the killer feature here. The deploy-wait-test-repeat cycle is what makes serverless development so frustrating compared to local Flask/FastAPI development.
I see others asking about CDK/Pulumi comparison - I think you're right that it's less about the underlying engine and more about the abstraction level. CDK gives you cloud primitives. Stelvio (like SST for JS) gives you developer workflows.
The automatic IAM permission generation is underrated. I've spent more debugging hours on Lambda permission errors than I'd like to admit. The error messages are terrible ("AccessDeniedException" tells you nothing about which permission is missing) and the docs always show overly-permissive examples.
Question: How does the dependency resolution work for native Python packages (e.g., numpy, pandas)? Are you pre-building wheels for Lambda's Amazon Linux environment, or using something like Lambda layers with pre-compiled binaries? That's historically been one of the most annoying parts of Python + Lambda.
8 comments
[ 4.4 ms ] story [ 30.1 ms ] thread[1] https://docs.aws.amazon.com/cdk/v2/guide/work-with-cdk-pytho...
[2] https://www.pulumi.com/docs/iac/languages-sdks/python/
[1] https://modal.com/docs/examples/hackernews_alerts [2] https://docs.ray.io/en/latest/ray-overview/getting-started.h...
edit: forgot to mention StackQL [3], which is somewhat similar but takes a SQL-first vs python-first approach. The IaC space is about to get a lot more interesting.
[3] https://stackql.io/docs
I see others asking about CDK/Pulumi comparison - I think you're right that it's less about the underlying engine and more about the abstraction level. CDK gives you cloud primitives. Stelvio (like SST for JS) gives you developer workflows.
The automatic IAM permission generation is underrated. I've spent more debugging hours on Lambda permission errors than I'd like to admit. The error messages are terrible ("AccessDeniedException" tells you nothing about which permission is missing) and the docs always show overly-permissive examples.
Question: How does the dependency resolution work for native Python packages (e.g., numpy, pandas)? Are you pre-building wheels for Lambda's Amazon Linux environment, or using something like Lambda layers with pre-compiled binaries? That's historically been one of the most annoying parts of Python + Lambda.