Open lann opened 3 months ago
That would be cool - but I'm not sure how we can sensibly emit those from the Shim in the way that the operator can reliably parse and proxy into the object - I wonder if @devigned has any ideas?
I think we should be able to pass that back as pod events via containerd.
More broadly, any terminal error which causes us to fail running a workload should be bubbled up to users in some coherent way. We should never leave a user to guess at the error which caused the failure of their workload.
Context: https://discord.com/channels/926888690310053918/950022897160839248/1275126957322801192 and https://www.youtube.com/watch?v=2CFj4_wt2qo&t=3609s
They were attempting to deploy a Wasm OCI artifact which is not yet directly supported by Spin.
Currently they just get the standard k8s "Deployment does not have minimum availability" status, but I wonder if we could give more useful feedback, especially when the app configuration has fatal errors.