Managed Ansible Automation Platform - Managed Active Node Counting

Overview

This article describes managed active nodes and the billing mechanics of Red Hat Ansible Automation Platform on Microsoft Azure.

Questions

Are managed active nodes defined differently between self-managed and managed Ansible Automation Platform offerings?

No. The managed node is still defined the same way based on the description in this article.

Managed offerings actively count managed active nodes and notify the host hyperscaler when a billing event for a managed node occurs, such as when a customer automates more managed nodes than they have purchased to automate within a billing cycle.

Self-managed offerings do not count managed nodes for billing operations at this time. But, Ansible Automation Platform and Ansible Insights will indicate how many hosts are in a customer's inventory.

Note: Ansible Automation Platform has active work in process that will improve the accuracy of counting managed active nodes in the future. This content will be updated to reflect those changes over time.

How are managed active nodes counted by managed Ansible Automation Platform cloud offerings?

A managed node is counted as a connection to a unique host (as used in the Ansible host variable) per billing cycle. Multiple automation events against this host will remain ONE managed active node for the billing cycle.

A unique host is determined based what Ansible uses to connect to the host such as the hostname or DNS record, IP address, or value in ansible_host variable. This is particularly important when the same host may be referred to in multiple ways.. Without the proper implementation, the same host may be counted multiple times.

The following table will represent an inventory. The "Host Variable" column is used to represent the host variable that Ansible uses to connect to the host. The "IP Address" column is used as a key to identify the same host referenced in different ways. The table represents two different hosts in total mapped to the two IP addresses.

Running an automation against this inventory would result in automating against four unique hosts since AAP would have used four different host variable values, even though three of those values point to the same host 10.0.0.0.

In this scenario, you have two options that can be used in order to ensure that the same host, 10.0.0.0 is not counted three times:

1. Update the inventory to remove duplicate references to the same host.
2. Leverage the ansible_host variable in the job template to refer to the same host using a common reference (IP, hostname, or FQDN will work as long as you use the same value across all hosts references to the same host).

---
ansible_host: vm.fully.qualified.domain

When using this approach, the number of hosts counted would be two.

Note: Microsoft Azure resource billing cycles are based on when those resources are deployed.

• Each customer has a unique billing cycle with Ansible Automation Platform on Microsoft Azure; start and end dates are based on when the customer deployed the managed application.
• Ansible Automation Platform Service on AWS billing cycles are all first-to-last day of the month.

How are managed active nodes counted when the entity automated is an API or control plane?

• Ansible Automation Platform connects to a host in order to perform automation operations. Customers control this configuration in their playbooks, but using a consistent connection scheme to “localhost” will mean that only one host is counted for any task that operates against an API or control plane.

• Each module in a collection is implemented independently, so you may need to verify with the collection creator/maintainer how the modules connect to resources. For example:

  • If you are using the Ansible collections for Azure azure.azcollection or for AWS amazon.aws, then only ONE managed active node is used for the billing cycle if automation is configured to run from localhost in each playbook.
  • If you are using an Ansible collection for any other vendor or control plane where the playbook instructs Ansible to connect to localhost, then only ONE managed active node is used for the billing cycle .

Note: Ansible Automation Platform has active work in process that will improve the accuracy of counting managed active nodes in the future. This content will be updated to reflect those changes over time.

Is it possible to limit the number of hosts automated during a billing cycle?

Ansible Automation Platform managed offerings do not provide any limit to automating hosts outside of the capabilities in Ansible Automation Platform, such as organizational host limits.

Is it possible to view managed active nodes metered?

Customers don't have the ability to monitor node count use when automating below their tier at this time. However, the article View Active Managed Node Meters can be used to create Azure notifications and alerting if an overage event occurs.