Summary
To determine the availability and health of each machine pool, we propose introducing heartbeats and status conditions of pools.
Similar to Kubernetes' node heartbeats & status conditions, there are 2 parties involved in this:
- The pool itself, regularly reporting its status by patching the *Pool resource's status and acquiring / keeping a lease object within the ironcore-machinepool-lease namespace. For references on leases, see the leases Kubernetes documentation.
- A pool controller within the IronCore control plane that updates a pool object's status conditions depending on the heartbeats of the pool.
It was decided to only implement this mechanism for MachinePools.
ToDos
Motivation
See the proposal's motivation section
Summary
To determine the availability and health of each machine pool, we propose introducing heartbeats and status conditions of pools.
Similar to Kubernetes' node heartbeats & status conditions, there are 2 parties involved in this:
It was decided to only implement this mechanism for
MachinePools.ToDos
ironcore-machinepool-leasenamespaceMotivation
See the proposal's motivation section