Chapter 8. Automation workloads
The default Ansible Automation Platform from GCP Marketplace is designed and licensed to automate 100 managed nodes.
8.1. Automation performance
Automating against your licensed managed node allocation in the offer is provided with the following operational expectations. Automating outside of the boundaries of these criteria is not supported, but can still function depending on your automation.
| Metric | Threshold |
|---|---|
| Concurrent Jobs | 10 |
| Forks per job | 10 |
Ansible Automation Platform from GCP Marketplace is driven using three n2-standard-2 instances, two of which run automation controller and one which runs automation hub. The automation controller instances collectively support a standard workload for 100 managed active nodes. Red Hat has tested this and proved that there is support for up to 10 forks each. This operating criteria has been set and tested using output-intensive, “chatty” workloads that produce 2 messages 7 seconds apart on both automation controller nodes. Workloads that are more I/O intensive might not work inside the boundaries of these conditions and can require the use of extension nodes to scale the deployment to support such automation.
8.2. Deployment scaling
If you want to scale your deployment beyond the initial number of supported managed nodes, Ansible Automation Platform from GCP Marketplace can be manually scaled using separately sold extension nodes.
Extension nodes are additional compute instances that can be deployed to scale-up or scale-out depending on the immediate scaling requirements. If your requirements are for higher parallel automation operations you can select compute shapes that scale-up, while if you have a requirement to automate more nodes over time you can select compute shapes that scale out.
Extension nodes are the supported way of extending the capabilities of Ansible Automation Platform from GCP Marketplace.
Red Hat does not support environments that are extended through customer design and implementation.