Hi,
We are looking to implement Snowbridge for an EKS deployment. What would you recommend to setup for a health check endpoint?
Snowbridge doesn’t have anything appropriate for this at the moment - but it has a relatively naive failure model which means it can get away with not having one.
TL;DR - in the circumstances where it would/should mark itself as unhealthy, it will just crash and k8s hpa will kill the pod and boot up a new one to replace it.
It’s not ideal I know! We hope to invest in improvements soon.
This would be a good feature to request in an issue on the repo, if you’re so inclined! (It is on the agenda to look into regardless, but there’s some work to do before we get to implementing this kind of thing).
1 Like