Recommendations and best practices
The following are some key recommendations and best practices to keep in mind when working with CloudWatch:
Create a monitoring plan for your infrastructure and abide by it. Note down all the metrics that you need to collect along with the method of its collection before actually deploying your infrastructure on the cloud.
Monitor each and every aspect of your infrastructure, including EC2 instances, EBS volumes, Elastic Load Balancers, and so on. Create specific alarms for monitoring each AWS resource independently
Avoid storing secret and access keys in your instances, and instead create and use specific IAM roles for permitting the instances to communicate with CloudWatch.
Create and check log files periodically for application or instance related faults and alerts.
Perform stress tests on your application and instances and create alarms that respond accordingly.
