AWS Well-Architected Framework: Reliability

Tal Shladovsky
June 16, 2023
min. read
Related Resource(s)
No items found.


The Reliability pillar of the AWS Well-Architected Framework focuses on designing and operating reliable and resilient systems in the cloud. This pillar includes principles and best practices on: 

  • Recovery from failures
  • Recovery testing
  • Horizontally scaling for high availability
  • Capacity planning
  • Automated change management


The Reliability pillar of the AWS Well-Architected Framework focuses on designing and operating reliable and resilient systems in the cloud. It involves building systems that can automatically recover from failures, scale to meet changing demands, and maintain availability in the face of disruptions. This pillar also emphasizes the importance of testing, monitoring, and continuously improving systems to ensure they meet the needs of the organization.

Design Principles

The Reliability pillar is based on the following five design principles:

  • Automatically recover from failure: This principle emphasizes the importance of building systems that can detect and automatically recover from failures, without the need for manual intervention. This includes designing for fault tolerance and implementing automated recovery procedures to minimize downtime and ensure that the system remains available.  
  • Test recovery procedures: This principle involves designing and executing tests to validate that recovery procedures are effective and can be relied upon to restore the system to a functional state in the event of a failure. This includes conducting regular drills and simulations to ensure that recovery procedures work as expected.
  • Scale horizontally to increase aggregate workload availability: This principle focuses on building systems that can scale horizontally to increase capacity and improve availability. This involves distributing the workload across multiple instances, which reduces the risk of a single point of failure and provides greater resilience and redundancy.
  • Stop guessing capacity: This principle emphasizes the importance of using automation and data-driven methods to dynamically provision resources to match demand, rather than relying on manual capacity planning. This includes using tools such as auto-scaling to automatically adjust resource allocation based on workload demands.
  • Manage change through automation: This principle involves using automation to manage changes to the system, including updates and deployments, to reduce the risk of human error and ensure that changes are made consistently and reliably. This includes using tools such as infrastructure as code (IaC) to automate the deployment and configuration of resources.

Best Practices

The Reliability pillar consists of the following four best practices:

  • Foundations: This best practice involves establishing a strong foundation for reliability by implementing core services such as compute, storage, and networking, and ensuring that they are configured securely and in compliance with industry standards. This includes setting up monitoring and logging systems to detect and diagnose issues, as well as establishing security and compliance policies and procedures.
  • Workload Architecture: This best practice involves designing and implementing workload architectures that are fault-tolerant, scalable, and highly available. This includes using services such as load balancing, auto-scaling, and fault-tolerant storage to distribute workloads across multiple instances and ensure that they remain available in the face of disruptions.
  • Change Management: This best practice involves implementing robust change management processes to manage changes to the system, including updates and deployments, to minimize the risk of disruptions and maintain availability. This includes establishing processes for testing and validating changes, as well as using tools such as infrastructure as code (IaC) to automate the deployment and configuration of resources.
  • Failure Management: This best practice involves designing and implementing processes to manage failures, including developing procedures for identifying, diagnosing, and resolving issues. This includes establishing runbooks and playbooks to guide response efforts, as well as conducting post-incident reviews to identify opportunities for improvement and incorporate feedback into future designs.

How can Stream.Security help improve AWS reliability?

  • With Stream.Security Discovery you get a full and clear picture of your cloud environment resources, their configuration and state, and the dependencies between them. This can be extremely helpful when needing to design and architect for high availability and business continuity of mission-critical applications or systems.  
Sample visualization of AWS topology, highlighting network resources and applicable rules/policies

Combining it together with Stream's Architectural Standards, to comply with security and availability best practices and industry standards helps you build a solid foundation for reliability.

Architectural standards provide a proactive and continuous way to implement best practices
  • Implementing a change management process is one of the Reliability pillar best practices. With Stream's Events you can track and be notified in real-time on cloud configuration changes of your cloud environment with the complete context and impact analysis.
    Track the who, what, where, and when of every activity in your cloud environment.
All activity logs get evaluated by architectural standards and relevant teams get notified of violations

In addition to that, Stream.Security Activity Logs provides build and runtime alerts for availability impacts with correlation to the problematic change.
Be the first to identify a downtime.

Activity logs provide a unified view to identify bottlenecks before you experience service disruptions
  • If you’re using Terraform to make changes to your cloud infrastructure, you should definitely use Stream.Security Simulation. Stream's simulation engine merges the current configuration state of your cloud in a combination with the Terraform code proposed change, to determine how your cloud is going to be impacted if the code gets deployed. It can be easily connected into your existing IaC deployment flow or run it as you develop in your favorite IDE (GitHub, Terraform Cloud, Gitlab, Bitbucket, Atlantis).
    Make config changes without breaking things.
    Deliver infrastructure as code with confidence.
Terraform simulations predict the exact impact of changes in your environment in terms of availability, security and cost


The Reliability pillar of the AWS Well-Architected Framework plays a crucial role in designing and operating highly dependable and resilient systems on the AWS platform. By adhering to the principles and best practices outlined in this pillar, organizations can enhance the reliability of their applications, minimize downtime, and ensure smooth operations even in the face of failures or disruptions.  

Key considerations for achieving reliability include fault tolerance, the ability to automatically recover from failures, and effective fault isolation. This involves leveraging AWS services such as Amazon EC2 Auto Scaling, Amazon RDS Multi-AZ deployments, and Amazon S3 cross-region replication to distribute workloads across multiple availability zones or regions, enabling redundancy and mitigating single points of failure.  

Monitoring and proactive alerting are crucial for maintaining reliability. By utilizing AWS CloudWatch, Amazon CloudWatch Alarms, and other monitoring tools, organizations can gain real-time visibility into their systems, identify performance bottlenecks, and take proactive measures to address potential issues before they impact end-users.  

Regular testing, such as load testing and chaos engineering, is essential to validate system behavior and resilience under various stress scenarios. Employing AWS services like AWS CloudFormation and AWS CloudTrail enables consistent infrastructure deployment and governance, while also facilitating auditing and compliance requirements.

Furthermore, by implementing automated recovery and backup mechanisms using AWS services such as AWS Lambda and Amazon S3 versioning, organizations can quickly restore services and data in the event of failures or data loss.

What's new