Update a cluster
The "Edit Cluster" Day-2 action shows the following form:
On the main tab, you could modify:
- Insecure Registries - (Optional) Modify the image registries which will only be accessible via HTTP. If you need to specify a custom CA for secure access to your registry then please follow the guide here.
- Cluster Owner Mail - (Optional) Modify the emails of the cluster responsible people.
- Custom Firewall Active - When enabled, allows to specify custom firewall rules.
- OIDC Enabled? - When enabled, allows to specify OIDC settings as specified here.
Click on the "Edit Cluster" button to start the update.
Restrictions
- If you only want to update the list of insecure registries, you still need to input the current value of workers in the worker nodes count field.
- If your cluster currently has at least one insecure registry configured, it is not possible to completely empty the list.
Worker Node Pools
The "Worker Node Pools" tab in the edit form allows you to manage the compute instances of your cluster. Please read more here.
Vertical Scaling
Vertical scaling is considered, as the modification that only involve existing pools and within that existing pools the attributes cpu, memory, storage, node-taints and node labels. To ensure a proper update of the cluster, where resource capacity of worker node pools is balanced out, we add an additional node for each worker node pool to have the same capacity available during the procedure, as the workers get drained one by one. After the vertical scaling operation is finished the additional extra node will be removed. Any other resource modification such as count adaption (up or down), adding / removing pool will lead to another path of execution and won't be considered as vertical scaling, thus won't trigger the creation of an extra node. Any other modification such as insecure-registries or email owner in addition to your resource modifications, will force the path to the normal update and you won't benefit from the extra node feature.
| scenario | Vertical scaling with extra node added | Normal scaling with ignore feature activated | Normal scaling without any protection |
|---|---|---|---|
| increasing count + modification of ( cpu or memory or storage or taint or labels ) | 🚫 | 🚫 | ✅ |
| decreasing count + modification of ( cpu or memory or storage or taint or labels ) | 🚫 | 🚫 | ✅ |
| No change to existing resources but adding new pool | 🚫 | ✅ | 🚫 |
| No change to existing resources (cpu/mem/.../labels) but adding more workers ( increasing count only) | 🚫 | ✅ | 🚫 |
| Adding new pool AND vertical scaling in other pools ( increase cpu for example) | 🚫 | 🚫 | ✅ |
| Removing 1 pool | 🚫 | 🚫 | ✅ |
| Removing 1 pool with vertical scaling in the other pools | 🚫 | 🚫 | ✅ |
| Vertical scaling in multiple pools | ✅ | 🚫 | 🚫 |
| Vertical scaling by changing node-labels | ✅ | 🚫 | 🚫 |
| Vertical scaling by changing node-taints | ✅ | 🚫 | 🚫 |
| Vertical scaling by changing storage | ✅ | 🚫 | 🚫 |
| Vertical scaling by changing memory | ✅ | 🚫 | 🚫 |
| Vertical scaling by changing cpu | ✅ | 🚫 | 🚫 |