Use upstream client tools to keep k8s cluster state in sync

piraeusdatastore / piraeus-ha-controller

High Availability Controller for stateful workloads using storage provisioned by Piraeus

Apache License 2.0

15 stars 8 forks source link

Use upstream client tools to keep k8s cluster state in sync #4

Closed WanzenBug closed 3 years ago

WanzenBug commented 3 years ago

Instead of manually running List+Watch using the basic API, we use the Caches and Indexers provided by k8s.io/client-go. They are better equipped to deal with temporary loss of active watches, which means we no longer crash every 10-30 minutes.