This article contains the list of known issues and limitations in Sitecore Managed Cloud Containers:
- Windows node pool is destroyed and recreated in production when scale in or out is performed. This behavior can be changed by applying the instructions from KB1001899.
- Windows node pool is destroyed during terraform rerun in cases of spot instances usage. For more details on the issue, refer to GitHub here.
- Application deployment needs to be rerun after each scaling of any Windows node. It is a requirement for disk space metric because the endpoint must be recreated for each node.
- For Sitecore XP production, the main node pool is recreated when an additional (scaled) node pool downscales to 0 nodes.
- NGINX send/read timeout cannot be set to more than 240. This limitation exists due to the Azure Front Door timeout configuration that cannot be more than 240 sec.
- Terraform might fail when removing the items from SKU config:
Provider produced an inconsistent final plan
There is a bug in Terraform that might occur if the solution has been deployed with a scaled node and when windows_scaled_vm_size and windows_scaled_node_count parameters have been removed from the configuration.
The solution is to set windows_scaled_node_count to "0" if a client does not require the scaled node.