AWS Cloud Resource Best Practices This article outlines two essential practices to help you avoid AWS Cloud service interruptions or exhausting your budget before you finish the program. The first practice is cleaning up the resources, and the other is creating concurrent resources within limits. Let's understand in detail. Stay within the Budget Udacity sets a fixed budget or credits for each student to complete their coursework in AWS Cloud Labs. Please understand that these credits are limited. Therefore, you must utilize the budget judiciously and only for educational purposes. To prevent unnecessary charges and efficiently utilize your AWS credits, we strongly advise you to delete all AWS resources immediately after use. You can always re-instantiate resources later as needed. Below are guidelines for deleting AWS resources to help you stay within the available budget or credits. EC2: Navigate to the Global EC2 view to view the list of all running instances from the EC2 dashboard and delete all EC2 instances across regions. VPC: You can delete custom VPCs, provided they do not contain ENIs, EC2, or NAT Gateways. RDS Databases: Delete your unused Database (DB) instances. However, before you do so, ensure that deletion protection is turned off. By default, deletion protection is turned on for a DB instance created with the console. Also, you don't need to create final snapshots or retain automated backups during deletion. S3 Buckets: Delete your S3 buckets if you no longer plan to use those buckets. Most importantly, ensure your S3 buckets do not allow public access to avoid data transfer costs. Redshift Clusters: Do not leave Redshift clusters running overnight or for longer for longer than necessary. You can shut down or delete your cluster to prevent it from running and incurring charges. CloudFormation Stacks: Delete CloudFormation stacks after you finish your exercise or the project. SageMaker Resources: SageMaker can quickly drain your budget. So, delete the models, endpoints, configurations, and notebook instances after use. Lambda Functions: Navigate to the Lambda console, select your function, and click Delete. CodeBuild Projects: Delete CodeBuild projects after use. Glue Jobs: Remove the Glue jobs and related resources. ECS Services: Delete the ECS services when not in use. EKS Clusters: Delete unused EKS clusters. In the EKS console, select your cluster, delete its nodegroup, and then delete the cluster. Bedrock: Delete the Bedrock Studio project when no longer in use. or Rekognition resources after use. Rekognition: Delete all Rekognition resources, including the custom labels model, project, and dataset. Don’t Hit the Concurrency Limits A strict limit exists on the number of concurrent (or parallel) instances and resources you can create at any given time. The limits set for these services are significantly higher than necessary to complete classroom exercises and projects. This policy applies to various AWS services, including EC2, CodeBuild, SageMaker, Lambda, Redshift, Bedrock, Rekognition, Glue, Fargate, and EKS. The exact limits vary by service and may change without prior notice. Exceeding these concurrency limits will result in the deactivation of your account and the deletion of your cloud lab resources. Therefore, you must keep your AWS account clean and regularly delete unnecessary resources to avoid hitting the limits. Conclusion Always clean up the resources and create concurrent resources within limits as prescribed in the exercises or the project. Adhering to these guidelines will help you effectively manage your AWS budget, avoid unexpected charges, and ensure uninterrupted access to cloud services throughout your learning journey.