Understanding configuration data

The SC//Reliant™ Platform Manager is a web console where you define the configuration of your SC//Reliant platform resources. It is a declarative-based solution where configurations, edge compute, and other settings are abstracted to data elements called Properties. There are two types of Properties: System Properties and Dynamic Properties. System Properties are common to all node environments whereas Dynamic Properties may be created as part of a specific feature or for a custom request. The SC//Reliant platform then uses Puppet to manage and deploy these configurations to your edge devices.

The following document explains how data management in SC//Reliant Platform Manager works conceptually.

Declarative configuration and blueprint-driven deployments

Configuration in the SC//Reliant platform is declarative and version-controlled. The platform uses Puppet as its configuration-as-code engine, with templates and modules written in the Puppet DSL and Hiera providing hierarchical data overrides. Configuration definitions live in source-controlled repositories and are applied through Puppet runs on each agent node, which means every site converges to the configuration you defined rather than to the result of ad-hoc local edits.

The SC//Reliant Platform Manager acts as the blueprint layer on top of that engine. Properties defined in the console at the site (tenant), node group, or node level become the inputs Puppet uses to render the effective configuration for each node. Changes promoted through the console flow through the standard release pipeline (for example dev → staging → pilot nodes → expanded rollout → full fleet), with SC//Reliant Platform Manager providing the targeting controls at each stage.

Multi-level inheritance and central governance

Real fleets are not homogeneous. You may have small-format sites, large-format sites, sites with regulated workloads, sites with specialty equipment, and the occasional one-off. The property hierarchy described below lets you express that variation without sacrificing central governance:

• Enterprise- and tenant-wide defaults are defined once and inherited by every node group and node.
• Node group properties override tenant defaults for a region, format, or function.
• Node properties override node group values for individual sites or devices.
• All values, at every level, are stored centrally in SC//Reliant Platform Manager and applied through Puppet runs that report convergence status back to the control plane.

Because every override is recorded in the platform and not in a one-off local edit, operators can view the effective configuration for any node in SC//Reliant Platform Manager, and drift detection and remediation run continuously: Puppet agent runs compare each node against its catalog and automatically reapply the desired state when divergence is detected.

Lifecycle controls at scale

The same property model drives lifecycle controls across the fleet. OS and security patches, container images, and VM images are packaged and distributed through the platform, and properties select which version each node group or node receives. That makes it straightforward to:

• Pin a pilot node group to a newer image tag while leaving the rest of the fleet on the prior version.
• Maintain per-node group or per-node maintenance windows so updates land during low-traffic hours.
• Roll back by reverting a property value: for container workloads, this redeploys the prior image tag. For VM workloads, it restores the prior snapshot.

The following sections describe the property hierarchy, the difference between raw values and property references, and the common use cases that arise when modeling configuration data in SC//Reliant Platform Manager.

Understanding the property hierarchy

A Property may be added to the node group, node, or both. The SC//Reliant platform prioritizes property values at the node level first, then at the node group level, and then at the site (or tenant) level. If there is no value set at a given level, then it looks for that data at the next level. For example, you can define your overall network structure for a physical location at the node group level, but override individual IP address property values at the node level and the hierarchy is respected.

When thinking about how to manage your configurations in SC//Reliant Platform Manager, you want to consider the following.

• When to specify data at the Node Group - For data that is common, and static, across all nodes you typically would want to specify that data configuration at the node group level.
• When to specify data at the Node - For data that is unique, or might change often, for individual nodes you would want to consider specifying that data at the node level.
• Lookup prioritization - When SC//Reliant Platform Manager is looking up the value of a property, it looks at the node first and then node group second. If a value is specified at the node level, that value is used. If no value is specified at the node level, it looks at the node group level.

Using raw values vs property references

All properties in SC//Reliant Platform Manager have a dropdown option for Raw Value or Property Reference. The following describes the differences between these two concepts.

• Raw Value - Specifying a raw value for a property means you're providing a specific value directly for the Property. For example, you can specify 192.168.0.0 as the value of an IP address property.
• Property Reference - Upon choosing Property Reference, the field will provide a drop-down list of available properties that have been added to, and are scoped for, the node group and node levels. Specifying a Property Reference as a value allows you to reference another property value to be used.

Examples

Take the following examples, which demonstrate how raw values and property references work.

When you specify a raw value of 192.168.0.1 for the Lan1 Ip Address property on a node, both the stored value and applied value on the node is 192.168.0.1. If you then use a property reference for the Docker_Access_Network property on the node group, which references the Lan1 Ip Address property, the stored value is Lan1 Ip Address but when the property value is applied on a node, it references the Lan1 Ip Address property value which is 192.168.0.1.

In the following example, you specify a raw value for a property at both the node and node group level. On the node group, you then use a property reference that references that property; SC//Reliant Platform Manager is going to ignore the value at the node group level due to the lookup prioritization and apply the value from the node.

Use cases

The following pages describe some common use cases that involve the data management in SC//Reliant Platform Manager.

Related documentation

• Managing configuration
• Node group property reference