Closed metaskills closed 1 year ago
using lamby with path overwrites today
This feels real far off the beaten path. Was that pun bad? lol
On a serious note, what does that mean? Most folks use Lamby like we do, simple SAM/HTTPv2 proxy and when issues happen we get reports. I'm not sure how or why folks would write their own integration outside of SAM but assuming that is what you're talking about? If so, I'd be more worried about folks that wrote custom code in their Rails app to work around this vs filing an issue. Does that make sense?
This feels real far off the beaten path
Sure, totally valid.
On a serious note, what does that mean?
Just pointing out that I could deploy a lamby app as a standalone with no http layer in front of it, and use a lambda proxy integration to invoke it from a decoupled gateway. If I were applying path overwrites in such a context, they would have worked in the previous release, but would break in this one.
I think it's probably ok to simply note that in the Changelog, just pointing out the possibility.
Thanks, I'll call out those in the changelog.
Here is an example of running local Rails under Puma with something like
/test/dwef782jkif%3d
in the path:Prior to this change, here is what it looked like in Lambda with Lamby:
We learned the same lessons with query params, always trust Raw. Thankfully, HTTP v2 allows that to happen.
ℹ️ There are some folks that use
AWS::ApiGatewayV2::Integration
with API Gateway to proxy requests to AWS Services like Lambda, but the mapping request parameter tooling (https://docs.aws.amazon.com/apigateway/latest/developerguide/http-api-develop-integrations-aws-services.html) does not provide a way to get to the raw value. RIP 🪦