Amazon VGT2 Las Vegas: Leveraging AWS AppConfig for Multi-Tenant SaaS Configuration Management

More Info

As a provider of Software as a Service (SaaS), adopting a SaaS operating model offers numerous advantages. One of the most significant benefits is enhanced operational efficiency, achieved by maintaining a singular software version for all tenants through dynamic configurations. This approach minimizes the need for multiple product versions and customization requests, allowing your team to concentrate on adding value to your offerings and fostering growth—the true essence of SaaS.

However, if you currently offer a SaaS solution, you might be facing obstacles in attaining your desired operational efficiency. You may find yourself responding to customer demands for bespoke customizations, resulting in a complex array of customer-specific code and pipelines that hinder your agility and detract from innovation and product enhancement. To overcome these challenges, you require effective tools and strategies that simplify complexity.

One effective strategy for improving operational efficiency is the implementation of SaaS pricing tiers. By categorizing configurations into pricing tiers, you can maintain a single software version while creating a diverse but well-curated tenant experience. Implementing these experiences based on pricing tiers necessitates a tool that can group application features, toggle them on and off, or set usage limits, all without modifying your application code.

In this blog post, we will explore how AWS Systems Manager AppConfig (AppConfig) can serve as a powerful tool for managing pricing tiers in a SaaS context. We’ll delve into AppConfig’s capabilities for managing pricing tier configurations and how it integrates with other AWS services. Furthermore, we will highlight how AppConfig can empower your development team to iterate and modify pricing tier features by simply changing configurations—without the need for code releases of your SaaS solution.

Multi-Tenant SaaS Pricing Tiers

Pricing tiers represent various subscription options or plans available to potential customers of your SaaS solution. The reasons for offering different pricing tiers can range from enticing customers to try your solution at a lower initial cost to appealing to distinct customer profiles based on industry and size.

From an implementation standpoint, a pricing tier can be viewed as a collection of features and usage limitations. Most SaaS solutions assign different features or varying limits on tenant-specific features to each tier. However, these groupings and limits are not fixed; as new features are introduced, it is essential to continually reassess your pricing tier offerings to ensure growth and success, particularly in the face of evolving market and economic conditions.

Given the advantages of pricing tiers, let’s explore how to implement them using dynamic configuration in AppConfig.

Introducing AppConfig

AppConfig allows for the management, storage, and deployment of application configurations, enabling consumption by multiple clients within your SaaS application. It is particularly useful for managing dynamic configurations like pricing tiers and can store a single JSON file containing dynamic configuration information, including feature flags.

While AppConfig is not the only AWS service capable of configuration management—services such as Amazon DynamoDB and AWS Systems Manager Parameter Store can also perform this function—AppConfig offers several advantages, including:

• Schema validation: Ensures configuration quality through Validators.
• Deployment strategies (including Canary): Defines how and when to change configurations.
• Integration with Amazon CloudWatch alarms: Automatically rolls back configurations if a change leads to application failure.
• Configuration versioning.
• Fully managed service.
• Compliance with multiple programs, such as the Federal Risk and Authorization Management Program (FedRAMP) High Authority to Operate.

Solution Overview

In the sections to follow, we will set up AppConfig to manage a feature exclusive to premium pricing tier customers. We will demonstrate how to fetch and evaluate feature flags within an AWS Lambda function, using the retrieved values in your code.

The flow includes:

1. A request from a premium tier customer reaching the Lambda function.
2. The Lambda function retrieves the AppConfig JSON configuration.
3. The Feature Flags utility evaluates the “premium” feature flag with the context of a “premium tier customer,” returning a value of “true.”
4. The Lambda function utilizes this “true” value in its code.

Let’s dive into the setup process.

AppConfig Configuration Setup

AppConfig comprises configuration hierarchies. Begin by defining an application—in this case, the ‘product-catalog’ application. Each application will have a list of environments (e.g., ‘dev,’ ‘stage,’ ‘production’).

Each environment encompasses the current configuration deployments, their statuses, and details. Furthermore, your environment can have multiple configuration profiles, detailing the deployed JSON contents and their current versions. We will explore these JSON configurations in the following section.

Implementing AppConfig Feature Flags in Code

After storing a configuration in AppConfig, the next step is to fetch the dynamic configuration containing the feature flags that define your pricing tiers. We will analyze how to utilize this configuration within a Lambda function to enable a feature for premium tier customers while disabling it for standard tier customers. The Lambda function must adjust its behavior in real-time, adapting to the input event and tenant tier during each invocation.

Using AppConfig’s Freeform configurations, a JSON document can be employed to outline the rules for our feature flags. While you can implement this logic as you see fit, we will utilize Powertools for AWS Lambda, a robust collection of serverless utilities, and the feature flags utility to retrieve the JSON configurations from AppConfig and assess our feature flags.

JSON Structure for Feature Flags

The JSON structure we will use includes:

• A default value under the default key.
• Optional rules that determine the evaluated value.
• Rules include a default return value (in case of a match) and a list of conditions.

For further insights on this topic, check out this informative blog post here. Additionally, for an authoritative perspective, you may want to refer to this source. Lastly, if you’re looking for a comprehensive resource on warehouse worker onboarding at Amazon, this review is highly recommended.