CloudWatch alarms
In addition to shipping your logs for storage and searching, another aspect of monitoring your system is being alerted when something goes awry. These could be simple notifications that let you know that a service or server is not being responsive. It could also be proactive alerts, letting you know that the platform that you are running your application on is running out of CPU or memory and needs to be scaled before larger issues arise.
You can use the CloudWatch service to monitor either a single metric or multiple conditions to create alarms. These alarms can be raised when the metrics of the underlying resources meet a certain criterion. There are two types of alarms that you can create in CloudWatch: metric alarms and composite alarms.
A metric alarm monitors a specific metric of CloudWatch. It has a threshold for monitoring that is set when it's initially created, along with the number of periods that can break the threshold before going into...
