We are close to putting Hubs into Kubernetes.

Why do I care?
1) proper K8s setup is easy to maintain and has great observability 
2) everyone on their mum runs Kubernetes

How did we do it?
@leewardbound PR'd into the helm chart repo 🫡

When is it live?
Currently testing sync + hub function.
Then iron out liveness, readyness & more
Chart works already though!

Full-stack digital product consultant @ BoundCorp.net - Regenfrogs.xyz - goldmongers.xyz - Reno NV

Developer 🤓 build, develop and operate Farcaster solutions @ dtech.vision also love tech, decentralization, fitness and oldschool rap 😎

Current heavy WIP version: https://github.com/dtechvision/onchain-helmcharts/blob/leewardbound-patch-1/farcaster/hubs/values-example.yaml

Code & 
Deploy | 
Testing | 
Testing | EU GMT+1 & 
Out of office at weekends

Autoscaling is a per user / infra configuration.

If you run multiple hub replicas then you may want to autoscale these up and down

Though you are unlikely to require more then 3 if we’re being honest

And even then yeah you have to make sure your hardware on the worker nodes can handle it

Then there should be no sync issues else Hubble will be improved for everyone 🫡

Humm. Will not cause hubble sync problems if it autoscales on demand.
As it happened last time the hub numbers increased?