AWS has recently introduced "out-of-the-box" best practice alarm recommendations for Amazon CloudWatch. This new option is designed to improve observability on the AWS platform, allowing users to easily add alarms from the console and download templates for CloudFormation, Terraform, and the CLI.
Currently available for 19 managed services, the new recommendations are designed to help engineers identify the metrics they should set alarms for and not miss significant events for their AWS infrastructure. Available as a toggle button on the CloudWatch console, the alarm creation wizard provides the metric name, statistic, and period filled in based on the alarm recommendation, following best practices for monitoring. If the recommendation includes a specific threshold value, this is also pre-filled.
According to the announcement:
You can now easily see all metrics with recommended alarms by applying an alarm recommendation toggle. We have made it easier to create alarms and now you can use the alarm wizard with pre-filled configuration for the recommended alarms. Alternatively, you can bulk download the auto-generated infrastructure-as-code for recommended alarms, so you can easily add them to your templates to automate your monitoring provisioning.
The services supporting alarm recommendations include Auto Scaling, S3, Lambda, CloudFront, DynamoDB, EFS, and Kinesis, with Amazon RDS, EBS, and SES being among the ones not yet supported. Allen Helton, ecosystem engineer at Momento and AWS Serverless Hero, comments:
This was a big gap in the observability space. I love that we can use this to tell us what we need to know about our applications.
It is possible to download infrastructure-as-code alarm definitions for recommended alarms and use the code to recreate them in CloudFormation, AWS CLI, or Terraform, with developers asking for CDK support.
Source: AWS documentation.
Shahed Kazi, global cloud architect at Everlight Radiology, comments:
This is a game-changer for logging and monitoring in AWS. Hopefully, in the future, AWS will provide simpler options to enable alarms in IaaS without the need for lengthy IaaS code.
To help customers understand the alarms, the cloud provider added as well in-line metric information to the AWS vended metrics displayed in the console, including the metric description, units, and statistics, and how they apply to the services. Corey Quinn, chief cloud economist at The Duckbill Group, comments in his newsletter:
In what can only be described as AWS finally lending you a helping hand, you can now get pre-packaged alarm recommendations for your favorite services. Nothing says progress like handing you an off-the-shelf list, rather than DIY your digital house of horrors. Congrats, CloudWatch – you're not as useless as you were yesterday.
The alarm recommendations and metric descriptions are available in all AWS commercial regions for 19 services, with more expected to be covered soon. While using the recommendations incurs no extra charges, the alarms are charged at the same rate as any other CloudWatch alarms.