This message was deleted.
# k3s
This message was deleted.
I think you might be exceeding the design goals of both etcd and kubernetes. If you’ve got so much data that etd itself is refusing to store it, simply switching the storage backend to kine+sql probably isn’t going to be a material improvement.
Sounds like perhaps some other rbac-enabled kv store or document store might be a better fit for your use case?
We wanted to take advantage of some of the benefits of the API server (provided API, version conversion of CRs, etc).
Is this amount of memory usage by k3s expected though?
Does k3s cache these objects as they are inserted?
not K3s specifically no, but Kubernetes does do a lot of caching yes.
You might try putting some memory limits on the k3s systemd unit to encourage it to be a little more conservative but at the end of the day you’re just kinda using it in a way it wasn’t really designed for.
It’s not meant to be a document/blob store. It’s an orchestration platform.
You might look at if all you want is CRDs without the orchestration bit?
it’s experimental though