The AWS Well-Architected Framework is a set of guiding principles designed to help dev teams build and operate robust, secure, efficient, and cost-optimized systems on AWS, or any cloud provider.
But why should you care about this framework? It's more than just a set of best practices; it's a strategic guide for building cloud solutions that are resilient, performant, and cost-effective. Adhering to the Well-Architected Framework helps organizations mitigate risks, improve operational efficiency, and ultimately deliver a better experience to their users.
At MultiTool, we believe progressive delivery is an underrated tool for helping achieve the best practices outlined as part of the framework, and ultimately a better experience for your end users. Progressive delivery, which includes strategies like canary deployments, ring deployments, and blue/green deployments, allows you to gradually expose new software versions to a subset of users. This approach minimizes risk and enables real-time feedback. Let's explore how progressive delivery aligns with certain pillars of the AWS Well-Architected Framework.
The Operational Excellence pillar focuses on running and monitoring systems to keep users happy while continuously improving processes and procedures. Progressive delivery is inherently aligned with this pillar because it emphasizes automation, real-time monitoring, and continuous improvement through iterative, low-risk deployments.
Automating the rollout of new features through canary or blue/green deployments reduces manual errors and ensures consistent deployments. Consider a CI/CD pipeline integrated with a progressive delivery tool: upon successful unit and integration tests, a new container image is automatically deployed to a canary environment. While you might have Prometheus metrics or CloudWatch alarms configured for elevated error rates (e.g., HTTP 5xx responses or increased p99 latency) on the canary instance, these threshold-based metrics are only half the battle. Once your alerts have fired, it’s already too late. In order to create a true feedback loop, completely automated rollbacks are crucial.
This loop, which must be driven by empirical data from production traffic, can also foster a culture of continuous improvement, as insights from each deployment inform future development cycles and refine automated processes.
The Reliability pillar is about ensuring a workload performs its intended function correctly and consistently, and additionally has the ability to recover from disruptions on its own. Progressive delivery directly contributes to reliability by:
The Performance Efficiency pillar focuses on using computing resources efficiently to meet system requirements and maintain that efficiency as demand changes and technologies evolve. Progressive delivery contributes to performance efficiency in several ways:
The Cost Optimization pillar is about avoiding unneeded costs and maximizing the business value of your cloud investment. Progressive delivery, while seemingly adding complexity, can actually lead to significant cost savings in the long run.
Consider the contrast between blue/green deployments and canary deployments in terms of infrastructure. While blue/green deployments offer excellent isolation, they often require maintaining two entirely separate, fully scaled production environments (e.g., two distinct Auto Scaling Groups and associated resources), effectively doubling your EC2/Fargate, EBS, ALB, and other resource costs during the deployment phase. Canary deployments, on the other hand, only require a small percentage of your traffic to be diverted to the new version running on a small compute unit, leading to a much lower infrastructure overhead during testing (e.g., spinning up only 10-20% of your typical instance count for the canary group).
Catching issues early with progressive delivery also reduces the likelihood of costly manual rollbacks, developer time building emergency fixes, breaking SLAs, and reputational damage that can lead to lost revenue in the long term.
Automated progressive delivery, with its focus on iterative, low-risk deployments, continuous feedback, and faster-than-human responses, seamlessly integrates with most of the six pillars. By embracing progressive delivery strategies, organizations can build more operationally excellent, secure, reliable, performant, cost-optimized, and sustainable systems.
At MultiTool, we simplify the implementation of advanced deployment strategies, providing the automation, observability, and control you need to confidently roll out new features, minimize risk, and continuously optimize your applications.
