You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
A worker pool (in the past we had just one, in the future customer can create many of them) represents a group of worker nodes. A worker node belongs always to one worker pool.
Worker pools have following attributes:
Name
HA- or Non-HA (this idefined by amount of zones: 1 zone=non-NA / 3+ zones =HA)
Machine Details
Machine Type (used for new created worker node)
Machine Architecture
Min and max amount of worker nodes (used by the node-autoscaler)
The following table shows how these information can be retrieved:
Attribute
How to find these information on SKR
Name
Each worker node provides the label ``. Aggregate all labels of all worker nodes to get the list of all available worker pools.
HA/Non-HA
Each worker node provides the label topology.kubernetes.io/zone (e.g. topology.kubernetes.io/zone: eu-central-1a). Aggregate all labels of worker nodes in the same worker-group to retrieve the amount of used zones within this worker pools: 1 zone = Non-HA / 3+ zones = HA
Machine Type
Could be retrieved from label node.kubernetes.io/instance-type (e.g. node.kubernetes.io/instance-type: m6i.large- maybe we also their show CPU/Memory size (all worker nodes are equal within a worker pool)
Machine Architecture
Could be retrieved from label kubernetes.io/arch (e.g. kubernetes.io/arch: amd64)
Min- Max amount of worker-Nodes
MIN-value = amount of used zones in a worker pool (a worker-pool has to have 1 VM per zone), MAX-value=TBC if we can show it - not available in SKR at the moment
Beside the nodes, it would also be beneficial to show per deployed workload (pod, deployment, job etc.) the underneath used worker-pools. Be aware that workloads with replicas can be distributed over multiple worker pools. The information of the used worker-pool can be resolved by checking which node is used by the pod and verifying the labels on the node.
AC:
Worker pools are visualised in Kyma Dashboard
Node overview shows the worker-pool a node belongs to
For each deployed workload (pod), the used worker-pools are visualised in Kyma Dashboard
Reasons
Introduce worker pools in Kyma Dashboard.
Attachments
The text was updated successfully, but these errors were encountered:
Description
By supporting multiple worker pools in Kyma, we should also visualise them in the Kyma Dashboard:
A worker pool (in the past we had just one, in the future customer can create many of them) represents a group of worker nodes. A worker node belongs always to one worker pool.
Worker pools have following attributes:
The following table shows how these information can be retrieved:
topology.kubernetes.io/zone
(e.g.topology.kubernetes.io/zone: eu-central-1a
). Aggregate all labels of worker nodes in the same worker-group to retrieve the amount of used zones within this worker pools: 1 zone = Non-HA / 3+ zones = HAnode.kubernetes.io/instance-type
(e.g.node.kubernetes.io/instance-type: m6i.large
- maybe we also their show CPU/Memory size (all worker nodes are equal within a worker pool)kubernetes.io/arch
(e.g.kubernetes.io/arch: amd64
)Beside the nodes, it would also be beneficial to show per deployed workload (pod, deployment, job etc.) the underneath used worker-pools. Be aware that workloads with replicas can be distributed over multiple worker pools. The information of the used worker-pool can be resolved by checking which node is used by the pod and verifying the labels on the node.
AC:
Reasons
Introduce worker pools in Kyma Dashboard.
Attachments
The text was updated successfully, but these errors were encountered: